[EM] When and how can we speak of "individual utility" and "social utility"?

[Abd ul-Rahman Lomax]

How to define individual utility in election methods is not 
necessarily a problem: the voter defines it. They system provides a 
means to express such utilities.

Aggregating utilities, however, is obviously not such a simple thing. 
But we should not let this distract us from the fact that utility 
analysis is really the *only* approach to judging how well election 
methods perform, it is not like we have other methods competing with it.

Election criteria might be considered such methods, but they are 
clearly indirect. Even the most basic of them, such as the Majority 
Criterion as usually defined, is clearly flawed in that we can easily 
propose election scenarios, and not rare ones but common ones, where 
it requires results that by any reasonable definition of election 
success are defective. As I've mentioned, if we can't use an election 
method for a group of people to pick a pizza, how can we expect the 
same method to work well with picking officers? Picking officers is a 
*more* difficult problem, not a simpler one.

Summing expressed individual utilities in order to maximize social 
utility is also defective, but the defect is on another level. And 
using more sophisticated methods of aggregating utilities may be 
politically impossible. Definitely, though, the topic is worthy of 
consideration.

However, we are working within, generally, the assumptions and 
constraints of democracy. And a basic assumption is that decisions 
should not be made against the wishes of a majority. There are very 
good reasons for this assumption, because the alternative is 
essentially oligarchy. The Majority Criterion would seem to be 
desirable, and, indeed, it is my position that the majority has the 
right of decision, *always*. However, this right is hampered and 
constrained if it is required that it be filtered through an election 
method that does not allow the majority to be fully informed before 
supposedly implementing its "decision." Indeed, I would argue, this 
is not a majority decision at all. If the Majority Criterion is 
satisfied by a single-stage election method involving choices among 
three or more options, the majority has not consented to the result, 
unless this is somehow made explicit.

Approval and Range do satisfy this, to some extent, but not 
perfectly. That is, where these methods fail to select the candidate 
preferred by the majority, they do so because the majority has 
consented, in some way, to that result. But, because of the need 
under many conditions of strategic voting, it may happen that this 
consent was constrained, and constrained consent is not consent.

The proposals that have been made that a Range election, where the 
Condorcet winner and the Range winner differ, be followed by a 
top-two runoff approach more directly the real majority right of 
decision. It would be more pure if the question presented in the 
runoff were a single question that can be answered Yes or No.

One of the problems in this area is that we assume that the majority 
consents to the election process. But unless the process itself 
includes a method for the majority, should it so wish, to withhold 
such consent, no election method is truly allowing majority rule, 
unless, again, results are ratified.

That we voted, perhaps, at one time, to use a particular method does 
not indicate present consent. Rule of law may be necessary (I 
certainly think it is), but rule of law, we should always remember, 
is not proof of satisfaction of the requirements of democracy. Rule 
of law, unless carefully monitored by an awakened citizenry, readily 
becomes oligarchy, wherever the law, intentionally or otherwise, 
establishes privileged groups.

To my mind, the legitimate answer to Warren's work is to develop 
better definitions of social utility, and methods of measuring it, 
not to simply pooh-pooh the whole idea because of imprecision of definition.

Sum of individual utilities is quite reasonable *if the majority 
consents*. The common bugbear raised up is the scenario where small 
but widespread gain is matched with high individual loss, or the 
reverse, small but widespread loss is matched with high individual gain.

The former can be tyranny of the majority, but the alternative to 
tyranny of the majority is tyranny of the minority. (Consensus 
systems attempt to avoid this, but simply substitute, unfortunately, 
tyranny of the minority where the status quo favors that minority. 
Consensus, properly, is a value to be pursued, and should not unduly 
constrain the majority as a binding restriction on it.)

And the latter is quite normal. We allow it and we even encourage it.

That Range voting equates the utility range of all voters is a 
protection against the extremities of these situations. To my mind, 
this approaches incorporating the majority principle into the method 
while still allowing the majority to consent to other than its first 
preference. It would be better, as I've said, if the consent is explicit.

Libertarians express some of this when they propose NOTA, None of the 
Above, as being standard on ballots. Approval with NOTA as a 
"candidate" and majority approval required for a winner satisfies the 
consent of the majority, about as explicitly as is possible in a 
single-stage method.

Range, similarly, with a NOTA candidate, and with the provision that 
if the NOTA candidate is preferred by a majority to all others, the 
election fails, we have a clear indication that to go ahead and 
select a winner is to violate majority rule.