Okay, one last post on this whole reservation issue and then tomorrow I swear I'm going back to fiction.
I continue to be dismayed by the (let's see, how shall I put this politely) obtuseness of arguments being made for IIT / IIM reservations. When I first posted about the issue on Sunday, I chose to go with a rant because enumerating the (to me) obvious reasons why the policy made no sense seemed redundant, and I was pretty clear that there was no reasonable justification to be given for so ludicrous a policy. Nothing I've read on the blogosphere in the last three days has changed my opinion on that, though in retrospect I realise that just because there's no argument to be made doesn't mean that people won't try to make it.
While there are plenty of inconsistencies and flaws in the arguments being made for reservations (most of them enumerated in dhoomk2's post here), the big problem is that almost all of these arguments are trying to argue the inequality is a problem and something needs to be done about it (which no one's really disputing) without really coming up with a coherent reason why IIT / IIM reservations are the way to go. That's like saying No, I don't have the wrong key, because look, the door is real. The issue here is not whether affirmative action should be taken to help the BCs at all, the issue here is whether introducing reservations in institutions of higher learning is the most a) effective (meaning will it really make a difference to those who are truly deprived) and b) efficient (meaning will the benefits from it justify the costs) of going about this. I'm yet to see anyone make a case for why reservations in IIT / IIMs are, specifically, the right way to go about addressing the problem.
You could argue, of course, that it's always worthwhile to do something - that any effort, however misconceived, to address inequality is a positive thing. That it's the thought that counts (see Amit's post here - which I entirely agree with). That's simply not true. That's like saying you should get points for answering a question in an exam even if you got it completely wrong. Public policy is not a buzzer quiz - policy actions like the reservations in the IITs / IIMs have real costs, so that decisions on them should be made after careful deliberation, based on hard evidence and after due consideration of all the alternatives. That's especially true if, as I argued in my last post, letting the government put the wrong policy in place means that they are under much less pressure to come up with the right one. Jumping to questionable conclusions based on no data and little logic will make things worse, not better.
A good way to think about this, I find, is what I call the "what you need to believe" test. Take the "why should we assume that people who get in under quotas will be poor students" argument. Let's try and imagine a scenario where that's not true, and where given that that's not true, reservations are the answer. Here's what you need to believe:
a) First, that on a systematic basis, the IITs and IIMs are not accepting candidates who would do as well or better than the students they currently take. Notice we're not talking hypothetical candidates who could have come up if primary school systems had been better - we're talking candidates who could join the institutes today and do as well if not better. This is not a marginal error, either, a full 50% of current students could be replaced without any drop (and a possible rise) in quality.
b) That these errors are strongly correlated with caste / backward class status - so that we don't need to understand what it is about the admission process that might be causing this error, what elements we might not be testing appropriately - all we need to do is let in 50% of candidates (selected, presumably, by the same process as the current batch, except with an additional constraint of being from a BC) and the error we've been making so far will automatically balance out.
c) We, being intelligent people and responsible citizens see why this step is necessary and have sufficient evidence to be satisfied that this step will work. Unfortunately, the authorities at the IITs / IIMs are either too dumb or too selfish (not that they gain anything by taking in students from upper castes, they're just selfish on principle) to see what we see and therefore must be forced into admitting students against their will.
See what I mean? The only argument for reservations in the IITs / IIMs I would buy (assuming you could provide evidence for it) would be one that said: the IIT / IIM admission process is actively discriminating against people from BCs (that is to say, they're picking people based on caste, not on merit), which is why we have to take action against it. But I'm pretty sure no one's arguing that - mostly because it happens not to be true. If anything it's hard to think of an admission system that's more meritocratic than the IITs / IIMs. And a large proportion of the professors I know back at WIMWI are involved in some social / socio-political project anyway. These are not near-sighted, unreasonable bigots we're dealing with, these are smart, concerned citizens. So if you manage to come up with a good argument for why reservations in IIMs make sense, the best thing you can probably do is take it to them and convince them to take it up.
More generally, though, I think the debate over this reservation issue speaks to the sad state of public debate on policy issues. The simple point is that public policy cannot and should not be based on rhetoric alone. What the government should or should not be doing, isn't just about principles - it's also about what will and will not work. If we're going to make critical decisions on important issues of national interest, we would hope that they were informed decisions - decisions that study the benefits and costs of available alternatives on an empirical basis. A number of people (on both sides of the reservation debate) have bemoaned the lack of empirical evidence on whether reservations work or not. I'm entirely supportive of that argument. I would love to see robust, carefully constructed empirical work that examines the real impact and costs of reservation policies while setting the right base lines . I'm not sure that work exists though (though admittedly it's not a field I've looked into in any detail - if strong evidence like that exists, why don't the supporters of reservations cite it, without getting into endless speech-making?), and in it's absence, rushing into a policy decisions that seem both counter-intuitive and potentially costly is rash and ill-advised.
To close, let me say this: my opposition to the reservation policy is not an existential choice. I don't see being opposed to reservations as a defining part of my identity. I'm opposed to reservations in the IITs / IIMs simply because I have good deductive reasons to believe that there will be a real cost to the institutes from the policy, while I've seen neither credible evidence nor convincing argument to suggest that it will help the truly underpriviliged.
If you still want to argue for reservations in IITs / IIMs, do this. Go to the Akanksha site. Look at the profiles of children there. Then come back and explain to me exactly how and why reservations in IITs / IIMs are the best way of helping these children or others like them (or, for that matter, how, short of a miracle, they're going to help them at all) - better than the initiatives I talked about in my previous post, better than all the things all the NGOs working in this sector are doing and could use support with. That's what genuine concern for social inequality looks like. Everything else is just crocodile tears.
 By which I mean that these studies need to control for what would have happened naturally, and must be statistically significant after correcting for random variation. Anecdotal evidence on how one or two people have done well means nothing.
 It doesn't even need to be evidence from India, btw - any evidence on forced admissions of students into top institutes would be acceptable, as long as there was enough detail to allow us to look into the relevance of example.