Tag: law of diminishing returns

The Value of Fatwas

With random ulemas here, there, everywhere (and maybe nowhere) issuing fatwas left, right and centre, I wonder if the value of the fatwa hasn’t gone down.

The thing with religion is anyone who is mildly religious will try to follow as much of the traditions and customs are possible. However, if one puts way too many restrictions, there is the chance that the follower might “do a ramanamurthy” * and just snap and decide to not any of the customs. AS long as you keep things reasonable, though, there is a good chance that the follower will continue to follow.

Now that the context has been set, I reiterate my question as to whether there isn’t a law of diminishing returns for fatwas. Things I suppose were fine when the fatwa was a rare entity. For example, twenty years ago when someone issued a fatwa to kill Salman Rushdie, it was a rare event (the fatwa) and hence got taken seriously and Rushdie has to go into hiding.

But look at the kind of fatwas that are being issued nowadays and I would be really surprised if these are getting taken seroiusly. For example, read this article (HT: Nitin Pai). There is a fatwa against buying insurance. There is a fatwa against working in banks. There is a fatwa against families accepting income earned by female members. And so forth.

Don’t the ulema understand that there exists a law of diminishing returns, and so people are not likely to take fatwas seriously if too many of them are put in place? Ok I suppose they don’t teach economics in Madrassas. Or is it that Islamic society is still in the part of the curve where slope is significantly positive ? (imagine a curve with the total “degree of acceptance” on the y axis and “number of religious restrictions” on the X axis. You would expect that the curve initially rises and then flattens out, and if you stretch things too far maybe even bend downwards).

All religions and all sects of all religions have their share of loonies. People who come up with random fundaes and then claim it’s part of the teaching of that particular religion and everyone should follow it. But I suppose that most other religions are decentralized enough that loonies are treated as just that, and people go on leading their lives without taking cognizance of the loonies.

PS: Check out this hilarious essay from The Dawn about this bunch of guys who tried to take along a Maulvi to Afghanistan to fight Russians.

The Eighty-Twenty Rule

I first got this idea during some assignment submission at IIT. One guy in our class, known to be a perfectionist is supposed to have put in 250 hours of effort into a certain course project. He is known to have got 20 out of 20 in this project. I put in about 25 hours of effort into the same project and got 17. Reasonable value for effort, I thought. And that was when I realized the law of diminishing returns to effort. And that was the philosophy I carried along for the rest of my academic life (the following four years).

The problem with working life as opposed to academic life is that the eighty-twenty formula doesn’t work. The biggest problem here is that you are working for someone else, while you were essentially working for yourself while you wree a student. Eighty was acceptable back then, it is not acceptable now. Even if you are working for yourself, the problem is that the completion-rewards curve is completely diffferent now.

Imagine a curve with the percentage of work done the X axis and the “reward” on the Y axis. In an academic setting, it is usually linear. Doing 80% of the work means that you are likely to get 80%. Fantastic. The problem wiht work is that the straight line gets replaced by a convex curve. So even to get an 80% reward, you will need to maybe do 99% of the work. The curve moves up sharply towards the end so as to give 100% reward for 100% work (note that I’m talking about work done here, not effort. Effort is irrelevant)

Now, why did I cap reward at 100% in the previous paragraph? Why did I assume that there is a “maximum” amount of wokr that can be done? Note that if there is a ceiling to the amount of work to be done, and to the reward, then you are looking at a payoff like a bond – the upside is limited – 100% but the downside is unlimited (yeah I know it’s limited at 0, but it is so far away from 100% that it can be assumed to be infinitely far away). Trying hard, doing your best each time, the best you do is 100%. But slip up a bit, and you will get big deficits. It is like the issuer of the bond defaulting.

Almost thirty years back, Michael Milken noticed this skewed payoff structure for bonds, and this led him to invent “junk bonds”, which are now more politely known as “high yield debt”. Now, these bonds were structured (basically high leverage) such that a reasonably high rate of default was built in. In an ordinary bond the “default expectation” is that the bond won’t default at all. For a high-yield bond, the “default expectation of default” is much higher than 0 – so there is a definite upside if the bond doesn’t default. So that balances the payoffs.

So how does that translate to work situations? You need to basically get yourself a job where there is significant scope for doing “something extra”. So that if you take into account the “something extra”, the “expectation” will be say something like 90% of the work. So by doing only a bit more than your old 80-20 rule from college, you can fulfil expectations. And occasionally even beat them, resulting in a major positive payoff (either in terms of money or reputation or power etc.).

The deal is that when the expectation is lower than 100%, the reward-work curve changes. It remains heavily convex for the duration within the expectation (so if expectation is 90% of work for 80% of profit, the curve will be highly convex in the {(0,90),(0,80)} area). And beyond this, it gets less convex and closer to linearity, and so gives you a bit more freedom.

I’m too lazy to draw the curves so you’ll have to imagine them in your heads. And you can find some info on convex curves here: http://en.wikipedia.org/wiki/Convex_function