Thoughts on a better system of presidential elections

[awfernan2002]

In my humble opinion, our current system of choosing our presidents would benefit from some improvement.  Before considering some of the alternatives to the present form of the Electoral College, let us review the goals of election reform.  Listed below are some of the important characteristics that are often cited as part of an ideal democratic election process. 

1.   One-Person, one-vote
2.   Election winner reflects popular mandate, avoid president elected with small plurality or minority of the vote
3.   Integrity of states’ rights
4.   Minimize election fraud and maximize transparency
5.   Expand opportunity for minor party and independent candidates
6.   Avoid complicated process that creates confusion and/or lowers turnout
7.   Campaign reaches throughout the country (adjusted for population density), not only a few select areas
8.   Winner determined by the people or through fair consideration on their behalf, prevent arbitrary decision by House or electors
9.   Votes convey as much information about voter preferences as possible

Indeed, it is quickly evident that many of these criteria are inconsistent with one another, and it is virtually impossible to satisfy all of them.  Thus, we must assign some relative weights to these attributes as we ponder the choices.

Popular vote
Votes are counted nationwide and the candidate with the most becomes president.  No more electoral college, end of story.  In my view, this alternative is unattractive.  Although it is to be commended for its simplicity and reflecting the direct will of the people, it could potentially result in a president elected with well under 50% support who would be vigorously opposed by the solid majority of the electorate.  Of perhaps greater concern is the fact that in the event of a controversial, close ending (which appears likely given the partisan show biz that has come to dominate the political culture), a “recount” would essentially have to be done nationally.  Similarly, to the extent that ballot stuffing or (on a less sinister note) weather patterns have an asymmetrical impact on turnout in different regions; the final result may be distorted.  Essentially, states forfeit their sovereignty in this process, as all 50 (plus DC) are blended together in a general pool.

Popular vote with run-off
In the event that no one has > 50% of the vote in the general election, the top two candidates would advance to a second round, to be held two or three (or whatever) weeks after the first general election.  Arguably preferable to the standard popular vote, as the president would by definition command a majority (not merely a plurality) of the popular vote in the second round.  Also, this system could allow for more clout by minor party candidates, as people could “vote their hearts” in the first round, while still allowing themselves the option to vote the “lesser of two evils” in the second round.  The major party (or two final) candidates would have to work to build a bigger coalition.  Note that the run-off concept could be applied in conjunction with the Electoral College on a state-by-state basis (e.g. Utah and Massachusetts give their EV’s to GOP and Dem candidates after first round, after getting over 50% of the respective states’ votes, while Ohio and New Mexico go into overtime).  Questions remain as to how much turnaround time should transpire between rounds (could election officials properly prepare?) and would there be voter fatigue or apathy?

Instant Run-Off Voting (IRV)
A more elaborate version of the run-off is IRV, whereby voters mark down their rank-order of preferences at the general election.  While still ensuring 50%, this method has the significant drawback of forcing voters to rank several candidates, many of whom they may not have even heard of.  IRV creates a large potential for confusion, and could result in paradoxes as voters achieve perverse outcome while trying to hurt their least favorite candidate.  No chance to weigh the likely effect of other voters.  However, IRV is innovative and perhaps useful in smaller-scale elections.  This system does have the significant benefit of conveying a large amount of information regarding voter preferences.

[awfernan2002]

Potential modified forms of the electoral college

To much better reflect the principle of one-person, one-vote, electoral votes could be apportioned by representatives, senators not included (i.e. 435 in total, or 436 with DC).  I tend to think that this would be a positive change, although it would somewhat detract from the “value” and autonomy of the states. This would be a global change, and one that can be applied to all of the various scenarios listed below.  In addition to how many EVs are allocated to each state, there remains the question of how many are awarded to the respective candidates.  I personally am not pleased that a statewide winner gets all of the electoral votes from that state; I don’t think it’s fair or representative for a candidate to get no credit for 47% or 49% of the vote.  A local election for governor has the same dynamic you may say, but that vote is all-encompassing.  In our unique electoral system, in lieu of a popular vote, each state is making a contribution toward the overall total, and I don’t think the contribution should be 100%-0%.  Plus, as has been mentioned ad nauseum, the campaigns have become so focused on only a few states, and particularly in the final month or two, completely ignore the solid majority of the country.


Non-winner-take-all formats:

Strict proportionality: Electoral votes allocated to each candidate in proportion to their vote share.  Methods to calculate this breakdown include largest remainder (Hare/Droop), or D’Hondt (which gives a small bias against small parties), Saint-Lague (similar to D’Hondt, but neutral).  Note that proportionality can be applied to a certain share of the electoral votes, not necessarily all of them, and could include a minimum threshold such as 2%, 5%, 10%, to make it less likely that a candidate would fail to get 50% of the electoral votes in a single round of voting. 

Other possibilities include combining a winner-take-all with proportionality.  How about designating 20% of each states’ EV’s to the statewide winner and the other 80% to be broken out in proportion.  Or, if you kept 538, you could give two to the state winner and then the rest done proportionally.  This will have the effect of making the small states close to all or nothing, while California is pretty tied to popular vote.  However, this arguably makes sense given the constituencies involved.

An important point that I believe has been overlooked in many discussions I’ve seen: the current number of 538 electoral votes would be very difficult to use for full or partial proportional allocation, as it would royally screw minor candidates.  If a state has 5, 6, even 15 electoral votes, candidates in the single digits will typically get zip.  More broadly, there is greater room for rounding errors, as a 4 EV race with candidate #1 beating #2 by a 60-37 margin could lead to a 2-2 split in the state’s contribution. 

Is it feasible to go to a 5,380 EV college, or 4,360?  How about 2,690 (i.e. 5 x 538)?  In terms of finding that many electors, I believe this is doable, as the Dem and GOP conventions combine for 6,000-7,000 delegates.  The EC hasn’t been changed much in several years - with population several times larger than 50 or 100 years ago, more electors is reasonable.

Another variation is to grant all of a state’s electoral votes to a candidate that receives over 50% statewide, while giving (semi) proportionally if winner is under 50%.

Smaller-scale winner-take-all formats:

Maine/Nebraska method: State winner gets two, congressional district winners get one each.  This is a a possible improvement that allows EVs to be granted on a smaller scale by a group that shares congressional representation.  Still I don’t think I like the double whammy of rigged congressional districts affecting both the House and the Presidency.  In some sense, it would be interesting to see how the re-districting process would be shaken up by this.  I think this is similar to the UK in that the score is determined by how many districts you win; big difference of course is that here, voters can split between parties.  In the current structure, there wouldn’t be that many close districts (within 10% either way, let’s say), so candidates would still be campaigning in select areas.

Another idea of mine is to allocate by county.  So, as an example, using 4,360 electoral votes nationwide (perhaps in addition to state winners getting 20 apiece), each county would get 1 EV per ~65,000 people or so.  To the extent that counties have fewer constituents, they would be grouped to approximate 65K.  States would still retain power as unique entities; specifically, the states would retain a fixed number of EVs, oversee elections, maintain the rules, have power to group counties (and perhaps change them at their discretion) and allocate a given number (or %) of electoral votes to state-wide winners (candidate would have thus incentive for big wins in big counties).  Counties make some sense given that they are already the unit of govt. where elections are organized and reported.  Given the smaller discrete units, it would be hard to coordinate a fraud that would carry a widespread impact.  Also, local communities would be casting their collective vote together, not people from East LA and Fresno, or Miami Beach and Pensacola.  However, we may still have the problem of most counties being lopsided; also, the processes of legislatures changing counties and voters moving between them could lead to some chaos.   

Beyond the question of how to allocate between states, it would be interesting to give the electors more leeway to vote as they pleased, following discussion and analysis after the election.  For example, if in 1992 Clinton got, say, 46% of the EC, Bush 41% and Perot 13%, the Bush and Clinton camps could make appeals to get Perot votes in order to try to avoid a House run-off before EC met in December.  If neither side was sufficiently appealing, the stalemate could force the matter into the House; Perot electors could stick by their man and send a message.  It hasn’t yet happened (at least not in a long time), but my guess is that having the House vote on the winner is an undesirable scenario (though, boy, it would make for great blog discussions!).  I think that having electors do so may be better than the House, as they would be picked by the campaigns/state parties, but could also weigh the views of constituents and discuss options with one another in a sober, possibly less partisan manner.  Importantly, they are independent of the political process, and are not part of the national legislative branch.

It’s clear that changing the system would be difficult and subject to intense political fighting.  If the Electoral College were abolished, or even if the number of electors were changed, this would require a constitutional amendment (I believe).  However, as things stand, states do have the discretion to allocate electors as they see fit.  Also, large changes would presumably have to be adopted by all or most states at once, as individual states’ clout could be diluted if they didn’t move in conjunction with others (e.g. Colorado considering moving to a proportional system).  Politically, the battle lines are not totally clear, but it would be probably be in the Democratic Party’s best interest to shift toward a popular vote.  Short of that, it is certainly the case that Dems would be better off if 435 (436) was the basis for the electors, as their candidate will win fewer states all else equal.  However, the blowout states do tend to be won by the GOP, so there would be a somewhat offsetting impact if a popular vote or proportional EV distribution was adopted.  Large state congressmen and other politicians of both parties should favor some shake-up to the system, particularly those from the Big 3: California, New York and Texas are currently getting significantly short-changed.

I didn’t really address elections for other offices, as the presidential choice is the big one being discussed right now.  In terms of presidential primaries, that could make for another long discussion in its own right and I won’t get into that here.   However, I like having a few small states vote early, with a staggered calendar versus a one-shot “national primary”.  It allows for voters to winnow out weaker candidates and to have a chance to see them in person before making a judgment.  Nevertheless, I think it’s important to space out the states farther, so that the impact of the first few is not too overwhelming.  Also, do Iowa and New Hampshire always have to be near the top?!

So, these are some of my thoughts.  I would love to hear some of your feedback and thoughts and what the future of elections should bring.

[muon2]

The Constitution constrains most of the alternatives.  Changing the number of electors or imposing a uniform national method of selection requires an amendment. National popular vote totals with or without a runoff also would require an amendment. It seems unlikely that an amendment could muster anything close to 3/4 of the states to ratify a change that would certainly weaken the many small population states.

For that reason, most action has been at the state level. This includes the recent failed CO referendum and the current attempt to bind enough large states to effectively create a winner by national popular vote. The result in CO shows the reluctance of a moderate-sized state to give up the potential importance of winner take all. As small, homogeneous states, ME and NE have less to lose by using a district based vote for electors. Large states would have as much problem with this as the small states do with direct popular vote.

A second drawback to any district system, whether by CD or by a more numerous set of districts is that the districts are prone to be gerrymandered by their legislatures. At present only the House is affected, but  widescale adoption of a system like in ME/NE would extend the effect of partisan state gerrymandering to a second branch of government. Gerrymandering districts can have more impact as the number of distrcits increase, so the big states would be most affected.

If the ME/NE system should gain popularity, there should be a mechanism to control the use of gerrymandering the CDs, particularly in the large states. This one area where Congress can impose controls. Much like Congress required the states to create single member districts of contiguous territory for the House, it could impose more structured rules on the creation of those districts. Rules about compactness and maintaining intact political jurisdictions certainly would fall in Congress' purview.