There's like, a bunch of interconnected issues in agriculture. For instance:
1. Our [read 'us' as being either 'my country' or 'human civilization'] food supply needs to be sufficient to provide adequate nutrition to all living humans.
2. Our food supply needs to be robust against failures in production, transportation and distribution which could be regional, widespread, and last several months or years.
3. Our food supply needs reach all individuals, so that no individual person lacks adequate nutrition.
4. Our food supply needs to grow at a sufficient pace to match population growth.
5. Our food supply needs to be 'sustainable' [used in the sense of 'our agricultural practices shouldn't lead to future shortfalls' rather than 'environmentally friendly'].
6. Our food supply needs to be 'environmentally friendly' [as in, minimizing pollutants, energy costs, land use, and damage to native plant and animal species].
7. Our food supply needs to be 'ethical' [as in, the moral rights of the individuals who participate it and the animals which constitute it need to be respected; what moral rights the people and animals possess is left as an exercise to the reader].
8. Small Farmer Proposition: Individual farmers as well as corporations and state-owned organizations should be able to participate in agriculture.
9. Mercantilism Proposition: Rich countries should not be able to crush the agricultural sector of poor countries through protectionism.
This isn't intended as a strict ordering of priority, but I personally care a lot more about making sure everyone is fed than some of the other issues, and I'll note that if we're not producing enough food, the inadequacy of our food distribution and feeding programs is kind of a moot point. My very-controversial opinion is that that famines are very, very bad and that we should heavily over-invest in preventing them. There are obviously limits to all investments, but it's one of the Four Horsemen for crying out loud.
Unsurprisingly, fixing all of the problems with agriculture is a complex problem, or else we wouldn't be talking about it because it would already be solved, and it's probably going to require many solutions, policies, regulations, technologies and programs to fix, and many of those are going to create more problems that will have to be addressed in turn.
But policy-wise, I think you really want the biggest, dumbest, least likely to fail solutions for the biggest problem of "no famines". Encouraging over-production creates other problems, like depressing wages for farmers and farm workers, both domestic and internationally, but it does a really, really good job of keeping famine at bay, at least domestically. Even in the ways it is prone to gaming or corruption, it still provides a more than adequate food supply. [This is a risk with provisions built on stockpiling; it is much easier for corrupt individuals associated with the program to just not stockpile and pocket the money, or sell stockpiles, or otherwise not do their job, and it can go on for years without being discovered, until you need those stockpiles and it turns out they never existed.]
So even if over-production creates other problems, or is "economically inefficient", I'd much rather see over-production continue, and additional policies, regulations and programs be added to address other concerns, both the problems over-production creates, and the problems it doesn't address, like the adequate distribution of food.