The problem is if a small group of MN owners are able to pass proposals that benefit themselves but not Dash -- skimming the cream off the top without spilling the milk. It's the way insiders and parasites have always operated.
The MN voting system is a terrific innovation, but the MNOs have to be vigilant, active and smart not to allow the system to be abused.
I believe that's the crux of
@camosoul 's critique -- that chunks of the Dash budget are being wasted and frittered away because MNOs aren't paying attention.
Jim & OP: I answered a similar question awhile ago over on BCT. I'll repost here, updated for current prices and MN count. RE: "How can we be sure that MN owners won't collude and skim budget funds if they have enough MN votes to do so."
This potential problem assumes that the vast majority of masternode owners just sit around and do nothing. Let's think about it:
a) Proposals have to be submitted about 96 hours before superblocks are created. It takes 72 hours for a proposal to "mature" and become valid, and then budget finalization happens around 24 hours before superblocks. Votes after budget finalization don't count. Consequently, we can say that masternode owners have a minimum of 72 hours to vote on any proposal.
b) An "abuser" would need to attract 10% net votes, which at the present time is about 410. Now, we're reasonably confident that nobody owns more than a couple hundred masternodes (assuming everybody is honest about how many they own). Even Otoh is down to 76 masternodes (again, according to him). So right off the bat, we're assuming that at least two people have to collude, and that's with ZERO masternode owners noticing the "attack" and voting against it.
c) An abuser (individual or, more likely, a group) would need to possess at least 410 masternodes, which is 410,000 DASH ($43 million at today's price). The maximum reward they could get from attacking the budget system is the entire budget amount, which is 6,920 DASH ($726,600 at today's price).
d) A successful abuse of the budget system would likely crash the price of DASH, at least temporarily. Let's say that such an attack reduces the value of DASH by just 10% (in reality, it would probably be far worse than that). The attackers gain AT MOST $726,600 but lose $4.3 million (10% of the value of their $43 million holdings).
e) In reality, there are generally many proposals that reach greater than 10% support. In fact, we often see the majority of the budget consist of proposals with greater than 18% support. That means the abusers would have to control even more masternodes to be able to seize the entire budget. In reality, only a small percentage of the budget is usually available to a "barely passing" proposal...usually less than 1,000 DASH. The most that an attacking consortium with $43 million worth of DASH could hope to seize is...about $105,000 (1,000 DASH * $105 USD per DASH).
Therefore, an attack on the budget system by a rogue group of masternodes is a) unlikely, and b) not economically feasible.
tl;dr Somebody who owns enough MNs to successfully abuse the budget system would lose far more money than he/she would gain.