[EM] Electoral College (was Re: Voting by selecting a published ordering)

Steve Eppley seppley at alumni.caltech.edu
Wed Apr 26 07:06:00 PDT 2006


Dave Ketchum wrote:
> On Sun, 23 Apr 2006 16:18:06 -0700 Steve Eppley wrote:
>> Some people don't consider the Electoral College winner-take-all 
>> within most states to be messed up.  Here are 2 reasons to prefer 
>> winner-take-all:
>> 1. If states allocate their Electoral College delegates 
>> proportionally, then every state would be a campaign battleground.  
>> The cost of campaigning would be much greater.
> HUH?  Being a safe state is a disadvantage:
>      The candidate who does not have to make any promises to help win, 
> can evade making promises to that state that could be a problem later.
>      The candidate who is sure to lose need not make any possibly 
> embarrassing promises to that state.

First, that argument does not appear to address my point in #1 above about the increased 
cost of campaigning if the Electoral College were allocated proportionally.

Second, I consider it an exaggeration to claim the candidates ignore the safe states.  For 
the most part, the issues the voters consider most important, and the voters' positions on 
those issues, have a lot in common in the safe states and in the battleground states.

Third, I'm curious how one can distinguish between these two cases:

    1.1  A candidate has a safe lock on some state, and therefore
         does not pay much attention to the policies preferred by
         the voters in that state.

    1.2  A candidate has a safe lock on some state, and has been
         paying so much attention to the policies preferred by the
         voters in that state that that is why it is a safe state.

>> 2. There would be an incentive to ask for recounts in all states.
> Not really.  You first consider which states tend to get different 
> counts after recount.  Then you consider which states are best prospects 
> for a profitable change:
>      Safe states:  Don't waste time on safe states, but invest in others 
> if worth the effort.
 >      Proportional:  Populous states such as NY are good prospects.
 > If a state such as Alaska or Hawaii just barely offered enough votes
 > for one member of electoral college, recount is very unlikely to
 > help - and could even lose that one.

I assume that by "safe states" Dave meant the states that are safe under winner-take-all, 
since under proportional rules it would be rare for a state to be so safe that a recount 
couldn't change the outcome there.  He and I appear to be agreed that there is no point in 
asking for a recount in safe states under winner-takes-all.  In winner-take-all states, 
the vote totals would need to be nearly tied for there to be a reason to believe a recount 
would have any effect.

What I wrote in #2 above was only about states that are proportional.

Don't forget that what can lose an EC delegate for one candidate can gain one for the 
other, so one or the other would have an incentive to ask for a recount there.  True, in a 
tiny safe state such as Hawaii or Wyoming, a recount is very unlikely to change the 
outcome.  So I should make a minor revision of my statement in #2.  Instead of writing 
there would be an incentive to ask for recounts in all states if all states were 
proportional, I should have written there would be incentives to ask for recounts in 
nearly all states, assuming the election was not a national landslide.

My point about proportional allocation leading to many recounts still stands.

I don't intend to dwell on the Electoral College.  It currently is not a big problem.  It 
doesn't interfere with the ability to make a significant improvement in how the President 
is elected, one which would lead to multiple candidates competing to be the best compromise.

>> What I propose the states do is tweak the winner-take-all formula so 
>> that instead of a sharp reversal when a candidate's total goes from 
>> 50% - 1 to 50% + 1, there'd be a linear change within the 49% to 51% 
>> region.  ...
> Looks like a loser:
>      Many safe states remain such.
>      Swing states make the two percent a BIG issue for more contentions 
> in recount, for gaining one vote on a near tie has 50 times the effect 
> of gaining one vote in the proportional world.

How could gaining one vote gain more than one Electoral College delegate, assuming this 
proposal were adopted?  Voter turnout would have to be near zero in some state.

What were the various counts and recounts in Florida in 2000?  The changes in the 
candidates' percentages were tiny.  It's implausible that a recount would gain more than 
one EC delegate in any state.

> Digging deeper, I see that Steve claims:  "With a formula like this, 
> recounts within a state wouldn't swing the state's allocation by more 
> than about 1 EC vote, so there'd rarely be an incentive to ask for a 
> recount."  Disagreed:
>      Conceded that likely profit from one optimum recount shrinks.
>      BUT, needing N gains, likely need to recount more states to try to 
> get there.

The number of states in which a recount might make affect the state's Electoral College 
allocation would be small in most elections.  Few states would be in or near the 49% to 
51% range.

The Electoral College totals before the recounts would have to be nearly tied, much closer 
than under the current system, for recounts in those few states to have a chance to change 
the outcome.

By the way, the range 49% to 51% that I proposed is not set in stone.  If someone thinks a 
smaller or slightly larger range would be better, I'd welcome hearing the reasons why. 
Also, perhaps it would be better if the size of the range depends on the size of the state.

> Note - Steve did recognize there could be more than two candidates, 
> requiring more detailed rules.

Not significantly more.  If I gave the impression that more than 2 candidates might share 
a state's EC delegates, that was unintentional, and I apologize for that.  Here's the 
general proposal:

    The percentage of Electoral College delegates awarded by
    a state to its leading candidate decreases incrementally
    from 100% to 50% as the lead decreases from 2% to a tie.
    The rest, if any, go to the candidate in second place.

More information about the Election-Methods mailing list