Regardless of whether you want to make the market as free as possible, or whether you want as many people to be as well-fed as possible, the long-run solution is the same for both [1]. The more the market is allowed to work, the better off people will be, and the more food they will be able to buy. Welfare systems offer a temporary solution at the expense of long-term growth, and so are overall negatives to making sure as many people can be as well-fed as possible. Importantly, efforts to artificially solve the high unemployment rates will similarly just increase poverty.
This reasoning is separate from the effects of welfare dependence. While they are also important considerations, they are hardly the only ones. The core consideration is as above: there exists a fundamental problem in the market, and throwing money at those most affected by it doesn't solve it, but instead just prolongs it and distributes the issue to those whose money is being thrown. In the long run, things get worse. India's growth is not helped by its programs, but hindered by them; its growth and its poverty levels would be better, not worse, without the programs.
So basically, there's no part of any of the calculus that involves ethics or anything. It's purely that we find, scientifically, that if you want the most people to be the most well-fed, market policies and low/no subsidies are the way to go. See: the Washington Consensus point 2 [2].
While that's all true, sometimes you have to accept some drag on growth to get from here to there. Market solutions take some time to settle out, and people who are hungry today aren't going to wait. Revolutions are a problem when you're trying to construct efficient markets, so as long as the growth arrow is going in the right direction it's probably wise to leave these kinds of programs in place. Presumably as the Indian economy modernizes they can be phased out.
This is a good point, but a contrasting point is that people who are hungry today aren't going to wait until tomorrow to get a job. It is incredibly valuable to be able to cut the benefits, so any cuts that can be made should be taken. Revolutions would happen if a dramatic change came about, such as the welfare programs disappearing overnight; but by market logic, the poor would not riot for welfare if the welfare had never existed, giving them no expectation of it, and giving them a lower likelihood of being poor in the first place. Chances are very high that these programs will never be phased out, and likely never cut in size over time--in the long term, welfare expands to fill the gap of how much the rest of society can afford to comfortably pay. But if they could be phased out, say over a 10, 15, or 20 year period, my judgement is that it could be sociologically workable. That is to say, if there is a level at which people switch from not rioting to rioting when their benefits are cut, you just cut near that level continually until you're close enough to find creative ways to stop sending checks altogether without them noticing.
India was a socialist country from independence, so that expectation has been there already for generations. I agree the best thing to do is phase out the benefits when you can. The trick is actually doing it.
This reasoning is separate from the effects of welfare dependence. While they are also important considerations, they are hardly the only ones. The core consideration is as above: there exists a fundamental problem in the market, and throwing money at those most affected by it doesn't solve it, but instead just prolongs it and distributes the issue to those whose money is being thrown. In the long run, things get worse. India's growth is not helped by its programs, but hindered by them; its growth and its poverty levels would be better, not worse, without the programs.
So basically, there's no part of any of the calculus that involves ethics or anything. It's purely that we find, scientifically, that if you want the most people to be the most well-fed, market policies and low/no subsidies are the way to go. See: the Washington Consensus point 2 [2].
[1]http://news.ycombinator.com/item?id=4449397
[2]http://en.wikipedia.org/wiki/Washington_Consensus#Original_s...