This proposal adjusts the maximum spending on proposals, in short words. The higher it gets, the less spending can potentially happen. The lower it gets, the more allowance it can permit. Although there is also a limit on the DHF itself, and depending on which proposals are voted on, that DHF daily limit can be or not all consumed. So, in a sense, voting for the right proposals is what matters. This slider (called the return proposal) acts more as a limitator of max spending and for the ones that don't want to vote for proposals that are at the top, as an expression of interest to say, we "don't agree" with those proposals being voted by the rest of the community.
This also acts as a defence mechanism, because lets imagine, that someone has 51% of the chain HP, and creates a proposal for getting DHF money for itself... then because it has 51% of chain HP, that person could upvote the proposal and automatically fund itself forever (assuming it keeps buying HIVE and continuously power up to be above or at around the 51%). Having a "return proposal" incentives others that don't agree with this, to stake slightly more HP (aka small cost), to refute the spending making the return proposal above the limit needed for the big 51% holder to keep. Eventually making the attacker go under 51% in stake, and decentralising the decision again.
Now, I know this is not perfect... because there is still cost to defend against this, but its relatively small cost. But we can potentially make it more efficient. Such as making the return proposal with a ratio multiplier... so that less people is needed for it to make an effect and for that ratio to not be linear... at low voting value, the ratio should be 1, but at high voting value, the ratio should increase to provide a method of "majority of the chain disagrees" and reduce the cost for expression of disagreement.
This can result in a problem too.. what if an attacker hates the top proposals, it can rapidly cease all the funding by holding majority of the chain HP... although, this sounds negative, this would mean that attacker is against where the DHF funds are going but he also can create a proposal for himself and get those funds himself... so, it self denies the ability to attack for self benefit.
Another attack could be the fact that someone wants to inhibit the chain development... by holding 51% of HP... well, if that person wants to do that, by holding at the same time 51%, then its a beneficial (for the community) attack, because keeping that 51% of HP will have to continue for a long time until the rest of the chain self balances, aquiring more HP. So, this would also contradict the act of attacking. And therefore, this aggressor would not eventually keep doing this, reducing his stake.
So, in a sense, I think a ratio could work... I should have probably made a normal post... instead of a comment. And please to anyone reading this HUGE comment... the examples I used here or numbers might not be technically 100% correct, but it was just to explore an idea. Anyone can comment... just brainstorming ok...