Arrow's impossibility theorem is not applicable and your bringing it up is meaningless nerdery. Providing value to society is not the same problem as selecting a single candidate from a set of candidates using a series of ranked preferences. For starters, value could be a function mapping preferences to a real number or random variable, called "utility," as opposed to individual ranked preferences.

(Also, even if it were relevant in any sense to talk about ranked preferences, they could still be consistently aggregable into a group preference -- Arrow's Impossibility Theorem merely says they might not be. It depends on what people's rankings are. Also, even if you replace the continuous set of options on how to provide value to society with a finite, discrete set of options, and then if you talk about a marginal penny of spending instead of spending at large, so that you can finally get people talking about a set of ranked preferences, you'll find that as you iterate elections of how to spend the marginal penny on a penny by penny basis, people's preferences change from election to election as laws of diminishing returns set in. For example, if you had five discrete option on how to spend money and assuming any reasonable degree of smoothness, such a series of elections would result in a spending distribution such that no simple majority thinks that the next marginal penny should be spent on a given option. )

I really can't understand how you could possibly bring up Arrow's impossibility theorem unless you're just trying to "win" an argument by bringing up some obscure mathematical reference that nobody wants to bother understanding so that they can respond to you. Because bring Arrow's theorem into a question like this is just flagrantly pure unadulterated bullshit.