Of modeling and predictions

Model of a system is not the system. It is an approximation, by definition.

When one models a system, certain assumptions are made, some aspects of the system is taken as invariable or at times independent of each other.

How then is one to determine if the approximation, the model makes sense?

The success of a model depends on its ability to predict. Models can be twisted and turned to fit the existing data and therefore can explain the known or the past in the most brilliant way possible. But if the model cannot predict the unknown or the future, then it is of no use. Einstein's model of the universe is an approximation but makes sense because it could predict certain phenomena that was till then unknown or unexplained. The success of environmental models will lie in its predictability. Unfortunately, we will get to know of the models' success only when it is too late.

I recently read (see here) that reduction in pollution will actually result in increased global warming. How? Pollution causes scattering of light which enables more leaves even those that are not at the top to carry out the process of photosynthesis at increased efficiency. Reduction in pollution will reduce the scattering and hence low photosynthesis resulting in low carbon absorption. So either you die of pollution or due to global warming! Nice!

The book I am reading now, The Logic of Life, by Tim Harford, is also about modeling. The premises are simple but effective. Human beings are capable of reacting to any situation rationally as well as irrationally. The book models the human society as rational and explores various non-economic issues. From what I have read so far, the rational model seems to be explaining the past well. There is no prediction of the future. I am hoping desperately that the book will make predictions somewhere down the line. But I won't be disappointed if there are none. Why? Well, if economic (and financial) models were that good at predicting, we wouldn't be facing the worst global financial crisis of all times, now would we?

Picture courtesy: Tory Byrne

Present and Immediate Danger

It took me a long time to figure this out but I have got it in the end.
I think.

Ok! Here's a question for you: Assume you know someone obese and he has a 80% chance of getting an heart attack in the next 10 years; and he is offered a medical insurance. Which insurance should he take?

a) Insurance that covers all his heart related disease.
b) Insurance that covers diabetes.
c) Insurance that covers all diseases.

If you have answered (a), then you are like the rest of us. Absolutely logical. But like most of us, wrong.
You see, (c) is a super set of (a) and (b) - (c) automatically covers (a) and (b).

What do you think the insurance man agent is going to push for?
You are right. (a). He may not even mention (c). Nor will the insuree - is that the word - ask about it.
The chances are that the premium for (a) is higher because the demand for (a) is higher.

We, humans, are wired to take into account present and immediate threat. That has saved us as a race for the last few million years. 'Present and Immediate' means we can see these threats coming. But this can also makes us vulnerable to hype. We can be made to see things that would otherwise not be apparent to us.

We can 'see' Global Warming coming, can we not?

We can see factories bellowing thick, black smoke.
We can see white, harmful smoke from the cars.
Our eyes sting when we stop at traffic lights and vehicles are idling.
We feel warmer every summer.

The obvious conclusion as repeated ad nauseum by magazines and newspapers is that the green house gas will cause global warming. And we are concerned.

We should be. Or should we be?

What about the good old concern about pollution?

In all this hype(?) about Global Warming, I rarely read any more about the impact of air pollution on our health.
What about the polluted rivers?
They don't cause Global Warming (or perhaps they do - I am sure some causal relationship can be figured out if you try hard enough) so you don't get to hear about that any more.

Tell me frankly. When was the last time you read an article in newspaper or a magazine the number of deaths due to pollution? Think hard.

Hmmm... sometimes, it is good to react to the 'present and immediate'.
I am all for systemic analysis (more on that in my blog, shortly) and take into account "in the long run" but for that we need to be alive in the long run.

Now consider the impact of taking away the present and immediate threat (pollution) and replacing this with a bigger picture (Gloal Warming).
Because pollution is no longer in our mind, we click our disapproval at all those people (always the other people) causing Global Warming when we read the newspapers, but do nothing about it.
We continue to air our disapproval but do not think twice while driving down to the nearest departmental store.

What do you think will affect us more?
a) A daily headline in the newspaper screaming number of hospitalisation and deaths due to pollution
b) A daily headline in the newspaper telling us about the possibilities of Global Warming in 20 years.


If you have answered (a), then you are like the rest of us. Absolutely logical. And correct.

Note: The picture used belongs to "jaylopez" (see gallery)