Options for Electoral College Reform

The recent presidential election has re-raised the question of electoral college reform, since the presumptive winner of the electoral college lost the popular vote, like Benjamin Harrison did in 1888.  Many are saying that if the loser of the popular vote serves as president, he will (like Harrison) be very weakened by a lack of mandate.  Therefore quite a few people, including Senator Hillary Clinton, are calling for a constitutional amendment that would elect the president by a pure popular vote.  In my youth I would have supported such a reform.  I remember Hubert Humphrey calling for it back in those days.  But there are quite a few other alternatives for reforming the archaic system we use, some of which might offer advantages over a simple popular vote.

The first question that has to be looked at is, What are the problems that we are trying to reform?  Obviously, we need to reform the habit of using cheap and unreliable voting equipment such as Votomatic card punches -- and not replace them with something even worse, such as a paperless touchscreen system with no real security -- but that's not a constitutional issue.  The concerns with the electoral college system itself are these:

1)  The candidate who loses the popular vote can win the election by being unpopular in the most populous states. This is not necessarily a problem.  The framers quite deliberately chose this rule, giving voters in sparsely populated states more weight than voters in heavily populated ones, so that the interests of smaller states would not be overwhelmed.  It was a compromise between backers of states (the existing power structure at the time, which many were reluctant to see weakened) and advocates of the people.

2)  The candidate who loses the popular vote can win the election if he happens to get small wins in many states while his opponent gets larger wins in fewer states, regardless of the size of the states involved.  This is because of the "winner take all" rule that most states use in choosing their electors.  It means that the proportion of the electoral vote often bears little resemblance to the popular vote.

3)  The small number of electoral votes causes a certain amount of random round-off error.  The winner-take-all rule makes this random error larger.

4) The winner-take-all rule also leads to voter apathy or disgruntlement in states where one party is dominant, because their vote will have no effect on the electoral vote totals.

5)  When no candidate gets a majority of electoral votes, the vote is settled by the House of Representatives, throwing out the people's vote entirely.  This generally leads to a purely partisan battle that loses all sight of whatever popular mandate really exists.  Often the only resolution is some kind of back-room deal like the "corrupt bargain" of 1824 or the similar mess of 1876.

6)  The electoral college tends to enforce a two party structure, freezing out alternatives, because nobody wants the election thrown to the House of Representatives.  Third party candidacies are generally seen only as "spoilers" instead of as real choices.  (The framers did not expect a two party system to arise; some cynics say they really intended to leave the choice up to the House of Representatives whenever nobody was overwhelmingly popular.)

7)  The forces upholding the two party system also bring about the necessity of primary elections, which have a host of shortcomings.  Or, if we don't have primary elections, the result is that most of the candidate selection process is done before voters have a voice.

8)  One problem with the primary system as it currently exists is: The parties are supposed to be private, independent organizations, not part of our legal apparatus of government.  Mixing the party's internal choice of candidates with the state election process is a bad compromise.  It violates private associations' right to choose their own candidates and platforms, and gives excess legitimacy to a side of the political process that doesn't deserve it.

9)  Another problem with primaries is that everything depends on the states that hold their primaries earliest.  States that vote late usually end up with no voice at all, because most of the candidates have conceded by then.  This leads to states constantly moving their primary dates backwards to get a more advantageous position, which in turn leads to the whole campaign season becoming more and more prolonged.

10)  The candidate that wins a party's primaries is often not the one who would best serve that party in the general election.  An ideologue tends to score better within the party than a moderate centrist does (though the current fad is for everybody to try to be centrist, since it worked so well for Clinton).

11)  You don't know anything about the individual electors you are voting to send to the electoral college, and in many states they are free to go against what the voters told them to do.  This creates an opening for a capricious individual to violate the voting rights of hundreds of thousands of citizens.

12)  The electoral college system -- and even more, the provision of the twelfth amendment that throws an election without an electoral college majority to the house of representatives -- makes it far easier to steal an election.  Instead of having to skew the outcome of hundreds of thousands of votes, as would be needed in a close popular race, an election stealing group can pick a few close states and tweak only a few hundred or few thousand votes.  This happened in 1876, leading to a series of several more elections of dubious legitimacy afterwards, and (depending on how you see things) either happened in 2000, or almost happened.

13)  And finally, the electoral college makes America look stupid to people from other countries, especially when we talk about how we're the bastion of democracy.

Now, what are the possible reforms?  Let's just go through a bunch of options we might try, and discuss the advantages and disadvantages of each one.  The first possibility is the most obvious and most frequently discussed one:

PURE POPULAR VOTE.  There are two variations to this one, depending on what you do when nobody gets a majority.  If you throw an election with no majority winner to the House of Representatives, you minimize the degree of change from the existing system.  If you give the presidency to the candidate with the largest plurality, you solve more of the problems on the list above, but you make it obvious how poor a mandate the winner has really got.  (The mandates of the people we elect now are no better, but the electoral college makes them look better.)

Advantages:  it brings us out of the stone age, it greatly simplifies things, it upholds the principle of "one person, one vote", and it gives us the opportunity to cut the House of Representatives out of the picture.

Disadvantages:  it does little to solve the problems above having to do with a two party system.  Indeed, it makes them worse in one way if the House is cut out: "split votes" can become a real problem.  If two similar candidates divide the vote of those who agree with them, then a dissimilar candidate gets the plurality of the votes even if the electorate sides more with the first pair on the issues.  Recounts in close races would be a real problem to implement.  But perhaps the biggest problem is that a constitutional amendment to elect the president by pure popular vote probably could never be passed, because an amendment requires approval by a large majority of states, and most states would lose power with this change.  Few of the less populous states are going to vote for a change that gives more power to the big populous states like California and New York, and less to themselves.  However, there are those who argue that the "winner take all" system actually gives California and New York more power already; if this is true and can be demonstrated to the satisfaction of the people in other states, then a popular vote reform amendment just might be possible.

POPULAR VOTE WITH RUNOFF ELECTION.  This is like the pure popular vote option, except that instead of holding primaries followed by a general election, you hold a free-for-all election followed by a runoff between the two top vote-getters.  This is the system commonly used for "non-partisan" elections such as mayor's races.

Advantages:  eliminates the bad compromises of primary elections, and reduces the control of entrenched political parties over elections, so that independent candidates have more opportunity.  This system reduces the split vote problem at least mildly; it is more difficult to figure out a scenario in which the candidate with the most popular views on the issues does not win.  This approach has a much better chance of electing centrist candidates instead of flip-flopping between the left and the right as partisan elections using primaries tend to do.  It shares all of the advantages (such as they are) of a popular vote with primaries, and it eliminates the house of representatives from the process.

Disadvantages:  it has the general disadvantages of any pure popular vote, though none of the specific disadvantages of a popular vote with primaries.  It would have the same trouble being approved by small states.  The clearest problem we've seen in contests of this type, such as mayorial races, is that there may be substantial difference in who votes between the free-for-all and the runoff; the outcome of the latter may not reflect the consensus will of those who voted in the former.  This is especially true given that only one of the two could be on the main general election day, and the other is likely to have a smaller turnout.  It would also be true if any great amount of time passed between the two elections, such as the current large gap between primaries and the general election.  However, that would probably not be as bad as the very severe split vote problem that would occur if you simply awarded the election to whoever got the plurality of a free-for-all.  That's why they have runoffs in the first place.  The split-vote problem is not entirely eliminated by a runoff, though.  If one party fields only two candidates who are similar in popularity, the other had better not field six or eight, or none might make the runoff.

PROPORTIONAL ELECTORAL VOTE.  This simply means that each state's electoral votes are divided in proportion to how the state's voters split, instead of allowing a winner-take-all system.  One or two states already do this.  We would have to be careful in exactly how we write the mathematical rule for how vote proportions are rounded off to whole electoral votes.  A variation would be to calculate each individual electoral vote according to the congressional district it represents, with the two senatorial votes being determined by the overall vote of the state.  This would leave the smallest states winner-take-all.

Advantages:  this is one of the few reforms listed here that might not require a constitutional amendment.  It could be mandated by Congress (though probably not if states protested strongly), or implemented one state at a time.  It would mean that the electoral college vote, within the limitations of roundoff error due to having only 538 votes total, would much more accurately reflect the popular mandate coming from the states.  It would preserve -- some would say restore -- the weighting of votes in favor of small states, which as noted above is probably very hard to eliminate, so we might as well consider that a good thing.  It would weaken the hold of the two party system: third parties would still tend to be seen only as spoilers, but their chances of breaking out of that role would be better.

Disadvantages:  if a significant third party effort is made, this reform would greatly increase the likelihood of the decision being made by the House of Representatives.  When that happens, your vote ceases to count.  And the variant of calculating votes by congressional district would make small states be winner-take-all while larger ones are proportional.

PROPORTIONAL ELECTORAL VOTE, PLURALITY WINS. This is like the above, except with the House of Representatives cut out of the picture unless there's an electoral vote tie.  Unlike the previous case, it would require an amendment.  Another factor that could be tossed in is an increase of the number of electoral votes -- say, ten or a hundred for each senator and congressperson, instead of one.  This would reduce roundoff error and make the proportional splitting of votes more accurate.

Advantages:  the result gets settled a lot easier and quicker without the House being involved.  This further improves the picture for third parties.

Disadvantages:  like a pure popular vote with the winner being awarded on a plurality, this makes "split votes" a real danger.  A conservative may beat two liberals even if the electorate has a liberal majority, or vice versa.  This leads any large power blocs to do their best to unify behind one candidate picked in advance and keep as much choice as they can away from the voters.

WEIGHTED POPULAR VOTE.  This is a system in which we don't use electoral votes, but still preserve the weighting in favor of small states.  There are various options for how to do this.  One is that we multiply each state's totals by a weighting factor which would be up to three times as large for the smallest states as for the big ones.  Another is that we could include a block of fake votes, of a total equalling 102/538ths of the number of real votes, and give each state an equal share of these fake votes, awarded winner-take-all to the most popular candidate in that state, or in proportion to the state's real vote totals.  As with those above, we have the choice of throwing races without a majority winner either to the House or to the candidate with the plurality, or holding a runoff.

Advantages:  this is very similar in effect to the proportional electoral vote option, only without the inaccuracy caused by rounding off to whole 538ths of the total.  If we use blocks of fake votes awarded winner-take-all by state, we reduce the likelihood of races having no majority winner.

Disadvantages:  using winner-take-all blocks of fake votes puts more of a freeze on third parties, but using weighting multipliers -- in effect, saying "Your vote counts 2.8 times as much as his vote" would create an overt impression of unfairness that, despite it being essentially the same as what we already have, would probably piss people off and create a lot of resistance.  Also, the "split vote" problem arises with weighting factors the same as it does in a pure popular vote.  The split vote problem is present, but reduced (rather artificially) with winner-take-all state blocks of fake votes.  A runoff would reduce it further.

ELECTORAL COLLEGE WITH POPULAR VOTE BONUS.  This one was proposed by Arthur Schlesinger Jr in Time magazine.  His idea was to award some extra electoral votes to whoever wins the popular vote.  The number he named was 102 electoral votes -- two for each state and for DC -- which is silly because it's just an overcomplicated roundabout way of awarding the election to the popular vote winner under all but the most bizarre circumstances.  But the question gets interesting if you award a bonus of less than 102 votes.  The amount you pick allows you to select any desired weighting between the outcome of a pure popular vote and the outcome under the existing system.

Advantages:  you can create a compromise, with any weighting you choose, between advocates of the traditional system and advocates of a pure popular vote.

Disadvantages:  like many compromises, there is very little to recommend this system, in any absolute way, over either of the pure alternatives it is compromising between.  Also, the process of selecting the exact weighting is bound to be contentious and arbitrary, and therefore repeatedly challenged.  It doesn't help that the exact degree of weighting for either side is not any obvious linear function of the number of bonus votes.

PREFERENTIAL POPULAR VOTE.  This is a system that was invented specifically to deal with split votes and lack of a true majority.  It is one form of what are called "instant runoff" balloting systems; if someone says "instant runoff" without further description, they are probably talking about preferential popular vote.  How it works is that instead of just voting for the single candidate you prefer, you make one vote for your first choice, a second vote for your next-best choice, a third vote for the next best choice after that, and so on.  If the candidate who was your first choice loses, your vote is transferred to the candidate you listed as your second choice, and if he loses, your vote is transferred to your third choice candidate.  The candidate who got the fewest first-choice votes is eliminated (his or her votes being transferred to the voters' second choice candidates), then the least popular remaining candidate is eliminated, and so on until someone has a majority.  Ideally, you rank every candidate from first to last, but this leads to a very complex counting process, so in practice we might ask people only for their first few candidates, not to rank all the lesser ones they don't like.  Limiting the number of rankings you vote for to three or four or five simplifies the logistics of counting while probably having only a small effect on the accuracy of the outcome, unless the number of candidates is very large.  This system is commonplace in Australia and New Zealand, and is also used in Ireland, with several minor variations by locality.  For instance, some local laws require that the voter rank every candidate on the ballot, whereas others allow the voter to mark only the first few that they like.

Advantages:  lots and lots.  No other system we've mentioned so far clears up as many of the problems listed above as this one does.  Split vote problems are eliminated because your second-choice vote goes to the other candidate who is on the same side of the issues; the candidate whose position on the issues is in the minority will lose even if the vote on the other side is divided several ways.  Third parties can thrive and attract exactly as many voters as they actually represent the beliefs of, because you have no fear of losing your vote to a "spoiler" if your second or third choice is a mainstream candidate.  This reduces apathy.  Primary elections become completely unnecessary: a party can run all its candidates at once and the strongest of the group will get almost all of that party's votes in the final count.  Elections in general will be driven more by people's preferences on issues than by loyalty to parties.  We could well see more candidates running as independents and being elected.  The election season could be significantly shortened.  No voter would be frozen out of the pre-selection process within a party because of variations in state laws.  The House of Representatives or other mechanism for resolving the lack of a true majority has no possible role except in a rare case where the vote is split so deeply that even people's third and fourth choices don't produce a winner.  (This is pretty much impossible where ideological camps are divided into Left and Right; it might happen if we had Left, Right, Up, and Down factions each incompatible with the other three, but that's hardly likely to happen.)

Disadvantages:  the ballot itself would be more complicated.  The counting process would have to be modernized and made a good deal more sophisticated and reliable.  (But then, the 2000 election makes it clear that we have to replace a lot of bad voting equipment already.)  We would probably need voting machines that have good preventive interlocks to reduce mistakes that would invalidate ballots, or there would be a lot more such mistakes made by voters.  (Again, this probably needs doing anyway.)  In any jurisdiction where we don't have such modernization, the counting process would be prolonged and tedious.  There would always be a tradeoff to make between having the voters make more secondary choices (fifth best, sixth best, and so on) which would improve certainty when there are lots of candidates, vs. limiting the number of secondary choices in order to reduce the data processing burden on vote counters.  (And if we ask for more secondary choices, a lot of voters probably won't make them; they'll vote for the two or three they like and cast no votes for the others.)  A disadvantage that has been claimed for this system is that there are obscure strategies by which clever enough voters can actually hurt a candidate's chances by giving him a higher vote, but I don't think this would apply if people were ranking only their top few choices.  It also might generate a lot of backroom horse-trading between various factions over who they will endorse as second choices, but I don't think that's necessarily a bad thing.  But as with a pure popular vote, one big disadvantage is that this undoes the weighting by states that the electoral college has, which means small states probably won't pass it.

PREFERENTIAL POPULAR VOTE POINT-COUNT (THE BORDA SYSTEM). This is a variation of preferential voting in which, if there are (say) six preferential rankings on each ballot, a candidate gets six points for being picked first, five for being picked second, and so on.  The candidate picked last gets one point.  The winner is the one with the highest total point count.

Advantages:  like preferential voting only with considerably faster and simpler counting procedures.

Disadvantages:  unfortunately, this system gives voters a considerable incentive to rank candidates who are threats to their first choice much lower than they would deserve based on the issues.  For instance, if voters for one mainstream party candidate want to make sure their man beats the other candidate from the same party, they would rank the other guy lower than the opposition party candidates.  If enough voters act this way, fringe candidates with no true mandate end up getting greatly inflated vote totals and could conceivably even win, a la Putney Swope.

WEIGHTED PREFERENTIAL POPULAR VOTE.  We could make the preferential system compatible with the electoral college's protection of smaller states by giving each state a weighting factor to multiply its vote totals by, as in the weighted popular vote system.  Or we could toss winner-take-all fake vote blocks by state into the mix, though I figure this would make an awkward and ugly fit with a preferential system.

Advantages:  all those of preferential voting, plus protection of small states as under the existing system, and therefore much better likelihood of being approved.

Disadvantages:  overtly spelling out in law that one person's vote counts for more than another's is bound to get people irked.  State winner-take-all blocks, on the other hand, introduce a non-preferential element, and increase the chances that the majority winner is not selected correctly according to people's real rankings of the choices.  As with any preferential system, the ballot and the counting are more complex and therefore would have to be more computerized than they are now.

PREFERENTIAL ELECTORAL VOTE.  This is an attempt to introduce the advantages of the preferential method into the electoral college.  The voters in each state would cast preferential ballots.  The state would award its electoral votes according to the first choice vote count.  If nobody got a majority of the electoral college, the votes of losing candidates would be transferred to those voters' secondary choices, and the electoral vote would be recomputed.  Electoral votes would have to be awarded proportionally by each state, not by winner-take-all, because if winner-take-all was used, the preferential part would quite likely never get to operate.  With a variation of this scheme -- have each state use a preferential system to select its electoral votes, instead of using the second place votes only when there is no electoral majority -- you could perhaps implement this system without a constitutional amendment (though the constitution does say that the power of deciding the voting system is given to the states, not Congress).  Even the first version might just possibly be able to be shoehorned into the space allowed by the present constitution, if the Supreme Court would allow each state to make its final choice of electors based on vote totals announced by the other states.

Advantages:  retains most of the benefits of a preferential system with less disruption of the status quo, and probably less contention over the state weighting issue.  Proportional electoral voting does not create the high risk of there being no majority winner as it would in a non-preferential system, if you use the stronger version that probably requires an amendment.

Disadvantages:  at best, this is not much more than an inaccurate way of doing weighted preferential popular votes; the electoral college apparatus is pretty much just nonfunctional window dressing.  At worst, the encumbrance of electoral votes and state blocs might spoil many of the advantages of true preferential voting for selecting the candidate who truly best represents a popular mandate.  If this is implemented in the toned-down form without an amendment, states might want to award their electoral votes winner-take-all, which would re-introduce a lot of inaccuracy into the process.

POPULAR APPROVAL VOTE.  This is a new addition to the list -- a system I had not heard of when I wrote the first version of this document.  Approval voting consists of giving a "yes" vote for every candidate that you can stand, and a "no" vote for all those you can't.  Simply put, you can vote for as few or as many of the candidates as you wish.  It's like preferential voting except without a hierarchy of individual ranking.  The winner is the candidate with the most total "yes" votes.  State weighting could be applied, or not, the same as with preferential popular voting, with about the same consequences.

Advantages:  the same advantages as the preferential system, and it's simpler, imposing no extra difficulties with counting or requiring new fancy voting machinery.  Split vote problems are eliminated, primaries and two-party constraints are eliminated, everybody can vote for who they really like best, and the one who is most broadly acceptable wins.  The final winner is almost certain to be the same person that would win a preferential vote.  At first blush it may seem that such a simplistic system cannot really replace the preferential method, but many who've looked into the system are confident that it's every bit as good.

Disadvantages:  the preferential system might do more to distinguish a ringing mandate from bare tolerance.  Or maybe not.  This gives less of a mandate on issues than a preferential system does.  Maybe.  One can't help but suspect that winning candidates will tend to be bland mediocrities... though the same may be true of any equally valid method of following a true popular mandate.  Still, voters might be happier and feel more engaged if they could indicate which candidate they really like vs. which they find merely acceptable.  Any really partisan voter might feel motivated to vote "no" for all candidates but one, thereby increasing the likelihood that the winner would have no majority.  In a widely divided race, the winner might therefore fall short of a majority, but that's a price you pay for any system that doesn't remove most candidates ahead of time.  In short, it's difficult to come up with anything very solid as a disadvantage for this system... all I've got here is either subjective, speculative, or trivial.

APPROVAL BASED ELECTORAL VOTE.  This is another one that could be implemented without an amendment.  The electoral votes of each state would go to whoever got the most approval votes in that state, or could be split proportionally among candidates according to their approval vote totals.

Advantages:  similar to those of the non-amendment version of the preferential electoral vote.  Voters would get to vote for who they really liked, and there would be no need for primaries and no obstacles to third parties.

Disadvantages:  Without an amendment, proportional assignment of electoral votes would leave considerable risk of nobody winning an electoral vote majority.  You could have ten candidates with handfuls of electoral votes apiece.  Winner-take-all assignment of states' votes would perpetuate errors and distortions of the outcome, without fully eliminating that risk.  A pure approval vote always has a winner, but combining the results of separate approval votes by state no longer has this advantage.  This could send third party candidates back to the ghetto of being "spoilers" once they become strong enough to win a few states.  A preferential system, if imposed in the more sweeping way that would require an amendment, could eliminate this problem even if the electoral college is still used; an approval system cannot do so.


In conclusion, I think it's obvious that I would strongly prefer either some kind of preferential system or a popular approval vote.  Any other leaves the majority of the current shortcomings unresolved.  I think a lot more voters would end up happy with the way they were voting, and we'd have far less apathy.  I think my preference of the systems listed here would be a pure preferential popular vote, or maybe a weighted one (though of course as a Californian I can hardly embrace weighting wholeheartedly)... but the longer I think about it, the more I lean toward an approval vote instead of a preferential one.  Third choice (and possibly the most obtainable) would be a popular vote with runoff, preferably with the two election dates only about 30 to 60 days apart.

I think the data processing challenges that a preferential system would bring are entirely manageable, even if voters end up casting many ranking votes.  Each precinct would, instead of submitting a total for each candidate, submit a table listing votes for each permutation of preference order.  The amount of data would be much larger than what is needed today, but would still be manageably sized -- a paper printout of it could fit into a manila folder if the voters rank the top eight candidates.  (It gets more like milk-crate sized if we allow nine or ten rankings.)  Once such tables are combined at a county or state level, the translation of secondary votes could proceed without any further reexamination of the ballots.  But if a precinct has inadequate data processing gear to produce and transmit these large result tables, they'd be forced to repeatedly recount all the ballots as losing candidates are eliminated one by one.  The burden could be minimized if voters only mark their top two or three or four choices, but this slightly increases the likelihood of the winner not showing a true majority.

One person objected that if we had a constitutional amendment describing such a system, it would be as big as the rest of the constitution.  I think it could be described in general terms in about the amount of text that the twelfth amendment uses, with the details being left up to congress or the states.  Another common objection seems to be that voters would be annoyed and confused by such a complex system.  But as far as I have been able to learn, when the system has been tried on a small scale by a few U.S. cities such as Cambridge MA, people are usually pleased with it.

I also think that almost any of the above would be better than the existing system.

If none of these national reforms gets accomplished, one positive step I'd like to work toward is to get the state of California (where I live) to allocate its electoral votes proportionally.  Having such an enormous block of votes be awarded winner-take-all is just far too unfair both to other states, and to the millions of voters on the losing side within the state.  It is, I believe, a major contributor to voter apathy.  Some may say that this would, at present, be to the advantage of the Republicans, but in many other times it would have helped the Democrats, so I think the idea can be considered on its merits for the long term in a nonpartisan way.  The fact that the state is genuinely split between left and right and is not dominated over the long term by either party (as New York and Texas, the other two most populous states, often seem to be) also means that such a reform has a real chance of being passed.  No state where the party in power always gains by winner-take-all would want such a change.  Unless maybe, just possibly, you could persuade northeastern democrats and southern republicans to make a trade, since the effect of both doing it together might come out pretty much neutral.  But California is the place to start.

If you have any feedback about these ideas, or if you have any reform concepts that I have not included, write me.
 


homeback to my home page

mailboxsend mail to Paul Kienitz