But, if all publishers gave feedback then the 'poor' authors would each save themselves from submitting to multiple publishers, and the 'poor' publishers deluged with submissions would instead have more refined manuscripts to evaluate and in lesser numbers. Thus, a little feedback can go a long way. This is because ideally manuscripts would improve only after a few submissions due to the feedback given and then they would either be accepted for publishing, referred to a more appropriate publisher, or shelved due to uninterested publishers.

Except for the fact that a large number of said 'poor' authors won't react that way. They'll argue, fight, and so on.