What It Is

I suppose I should explain the basic mechanics of the national popular vote campaign. The animating idea is to take advantage of the fact that states have, under the constitution, essentially unlimited authority to allocate their electoral votes however they want. NPV encourages states to pass laws saying that their electoral votes will be allocated to the popular vote winner if and only if enough states to comprise a majority in the electoral college also adopt such laws. In short, when a state passes an NPV law nothing happens at all in the short term. But if enough states do it to "win" 270 electoral votes, then suddenly there's a gestalt shift and the country has, in effect, a popular vote system.

It's a good idea and it's made a lot of progress in recent years. Chris Pearson reports, however, that somewhat oddly this has become a partisan issue at the state level with Republicans usually in opposition. This he plausibly speculates is the legacy of 2000 where people see NPV as not much more than an effort to degrade the legitimacy of the Bush presidency. In fact, though, this should really just be seen as something that pits state against state. The currently "uncompetitive" states -- places like California and New York and Vermont and Massachusetts but also places like Texas and Alabama and Mississippi and Utah -- are the ones mostly clearly disadvantaged by the current system and that goes for Democrats and Republicans alike.

Comments

It strikes me that you would want to set the trigger a little higher - imagine if all the blue states + a few swing states pass this thing, and then you dilute CA, NY, IL etc. without diluting TX, AL, MS, etc. I don't know where I would set the trigger, but not at 270.

Posted by minderbender | July 18, 2008 11:56 AM

I'm going to keep bringing this up. Popular Vote rules will mean that winning by 500,000 votes in California or Texas will be more important than losing by 50,000 votes in 9 other states. It will ENORMOUSLY tilt Presidential elections toward running up the score in big base states and will mean that no small state will EVER AGAIN play any significant or meaningful role in ANY way, because the margin of victory or defeat in small states will never be big enough to practically affect the popular vote total.

Mike

Posted by MBunge | July 18, 2008 12:01 PM

Republicans are against it because they represent the non urban areas right now. Not sure why you find this to be a surprise.

Posted by Rob | July 18, 2008 12:05 PM

Mike's argument is interesting, but it misses the point a bit. The real effect will be to concentrate campaigning where each vote is cheapest (in time and money), which presumably means dense areas. State borders won't matter anymore, so it's true that candidates will want to spend time where the voters are, but this doesn't mean big states per se.

The other complicating factor is that electoral votes are discrete, so possible no one would ever campaign in a two-district state, because it would take a huge margin to split it 3-1.

Posted by minderbender | July 18, 2008 12:08 PM

MBunge, the problem is that small states are basically "playing the lottery" that they hope that, if they're lucky, they MIGHT get to play a significant or meaningful role.

Seriously, when was the last time presidential candidates cared about North Dakota or Vermont?

On the other hand, New Hampshire seems to have won out in that game in recent years.

Everyone paid attention to Ohio and Florida, but no one seemed to care much about Indiana.

I see what you're saying, but, really, I don't see that the possible flaws of the proposed system are any worse than the flaws of the current system.

Posted by Tyro | July 18, 2008 12:09 PM

MBunge:

Here is your campaign schedule w/o the EC:

Day 1: Boston am - New York pm
Day 2: Philly am - DC pm
Day 3: Atlanta am - Miami pm
Day 4: Dallas am - Houston pm
Day 5: LA am - SF pm
Day 6: Portland am - Seattle pm
Day 7: Detroit am - Chicago pm

Repeat every week from July to November.

Posted by Blue Moon | July 18, 2008 12:10 PM

Electoral votes are slightly tilted toward less populous states, which are more likely to be Republican.

Posted by Al | July 18, 2008 12:12 PM

I think this campaign is pretty interesting, but I do worry about the legal challenges that might result if this ever went into effect. The issues that would be raised by this legislation would be unique and would require some judge to make new law to resolve. There are a lot potential for things to go wrong in that process, particularly if the ruling might change the result of a presidential election.

Posted by Leo | July 18, 2008 12:17 PM

MBunge, the flip side of that is that you won't have situations where parties ignore big states, like Al Gore did with say, Texas; or Bush did with New York or Illinois. What you're really saying, Mike, is that Presidential elections will be ENORMOUSLY tilted towards voters, rather than the current system, which is ENORMOUSLY tilted towards land.

Posted by Dr. Victor Davis Handjob | July 18, 2008 12:18 PM

"State borders won't matter anymore"

As long as state governments exist as distinct political and governmental entities, state borders will continue to matter quite a bit. Unless you conceive the Popular Vote business as a disguised effort to promote the return of the city-state model of public life.

And I don't "cheap votes" necessarily mean dense areas. It will refer to areas where spending the least amount of money results in the biggest vote gain, which will likely mean places that are heavily dominated by one side or the other.

Take two population centers of 500,000 people. One is 52/48 Dem v. Rep. One is 60/40 Dem v. Rep. If you're a Democratic campaign with a million dollars to spend, do you spend it equally on each area?

Mike

Posted by MBunge | July 18, 2008 12:19 PM

"What you're really saying, Mike, is that Presidential elections will be ENORMOUSLY tilted towards voters, rather than the current system, which is ENORMOUSLY tilted towards land."

No, the current system is ENORMOUSLY titled toward states, distinct and separate political entities that play fairly important roles in American public life and governance.

Mike

Posted by MBunge | July 18, 2008 12:28 PM

On Yglesias' main point, the analysis of who benefits only makes sense if the main goal is getting candidate money or attention. That makes sense in the case of pandering. Florida's sugar cane industry probably benefits from the importance of Florida in national elections.

On the otherhand, conservatives in competitive states may think that the advantage they get from the overrepresentation of small conservative states outweighs the advantage of getting flooded with campaign ads.

With regard to MBunge's point about voting splits, I think Yglesias had a post on this point a bit back, but the political split does not actually tell you which votes are most changeable. These are mostly two different kinds of factors. There may be more voters who could change in a candidates direction, but it would be harder to actually get them to change. Or a region could appear more friendly because ones popularity is higher, but that is because one has already gotten all the votes that lean in ones own direction.

Given that the denser markets tend to be the more expensive markets, it probably is not as simple to guess what markets will get the most attention. From a strategy perspective it actually throws things more widely open.

Posted by Lon | July 18, 2008 12:38 PM

It strikes me that you would want to set the trigger a little higher - imagine if all the blue states + a few swing states pass this thing, and then you dilute CA, NY, IL etc. without diluting TX, AL, MS, etc. I don't know where I would set the trigger, but not at 270.

I think you've misunderstood the mechanism -- if 270 EVs worth of states pass this thing, then all those 270 EVs go to the popular vote winner, and that person gets the presidency. It doesn't matter how the rest of the states allocate their EVs.

What worries me about this is what happens in extremely close elections; a recount could be a horrific mess, especially if the legal triggers for a recount in one state depended on the counts of votes in another state.

Posted by Matt Weiner | July 18, 2008 12:38 PM

Well, of course the DEMOCRATIC party would benefit from having a DEMOCRATIC election!

Who cares if a nascent war-torn republic like Iraq can have higher turnout for direct national instant run-offs? They're too complicated so we can never do them here!

Oh, and creating federal voting rights would undermine the Jim Crow laws that have been giving Republicans victories in the South all this time.

Posted by Luke | July 18, 2008 12:41 PM

Mike,

First, in the current system candidates typically ignore plenty of states. That is actually a consequence of the winner-take-all approach--if you have little chance of winning/losing a state, you have little incentive to campaign there.

Second, states will remain part of the conversation insofar as they remain relevant political units, but each individual state would revert to having around as much weight as their population dictates. Which seems pretty fair.

Third and finally, it isn't at all obvious where your marginal voters might be. They could be in districts you are likely to win anyway, or in districts you are likely to lose anyway, or so on, and it ends up being quite likely that every state would have some of your marginal voters somewhere.

In fact, Obama's primary campaign under a proportional system was good evidence of this. Except for the Super Tuesday state, when he was somewhat resource constrained, he found places to campaign for marginal voters in every contest, even the ones he knew he was going to lose statewide. And even for Super Tuesday he campaigned in a lot places, including states he knew he would lose and states he knew he would win.

So really all that is going on here is that a few states (some big, some small) are protecting their disproportionate influence on Presidential politics at the cost of the rest of the states. Indeed, that is why as FairVote related, in 1969 the House of Representatives, with support from both LBJ and Nixon, voted overwhelmingly for a direct election. And a majority of the Senate supported it too, but it was blocked by a filibuster. So, it really is just a minority of states thwarting the rest.

Posted by DTM | July 18, 2008 12:47 PM

"No, the current system is ENORMOUSLY titled toward states, distinct and separate political entities that play fairly important roles in American public life and governance."

Except when they're not competitive, like California, New York, Texas, or Illinois. Then one of the two parties is ENORMOUSLY LIKELY to write them off completely in the general election. Thanks, Electoral College!

Posted by Dr. Victor Davis Handjob | July 18, 2008 12:51 PM

It will ENORMOUSLY tilt Presidential elections toward running up the score in big base states and will mean that no small state will EVER AGAIN play any significant or meaningful role in ANY way, because the margin of victory or defeat in small states will never be big enough to practically affect the popular vote total.

Only true so long as the marginal cost of racking up another vote in a large state is less than gaining another vote in a small state.

Suppose that candidate A is projected to get 8.5 million votes in California in high turnout (around 60% of the vote) and 750,000 votes in Iowa in moderate turnout (around 50% of the vote). The formula for determining the cost of winning another vote net (CAV) in each state would be (CAV = PVA * PNVG * CC), where PVA = (1/probability of targeted potential voter actually appearing), PVNG = (1/(probability of targeted potential voter actually voting for you - probability of potential voter actually voting for your opponent) and CC = campaigning cost per voter in that state. Since California's turnout is already higher, raising turnout in California requires targeting more marginal voters than in Iowa, so let's say that for this case PVA(iowa) is 1.4 (70% of the target group shows up to the polls), and PVA(cal) is 2.5 (40%). Since more marginal voters tend to be low-information and are less likely to have a strong partisan preference, let's say that PVNG(iowa) is 1.7 (your candidate wins 80% of the target group), and PVNG(cal) is 3.3 (65%). PVA(iowa) * PVNG(iowa) = 2.38, while PVA(cal) * PVNG (cal) = 8.25, or 3.47 times the equivalent figure for Iowa. Since California has about 12 times the population of Iowa, it's better to devote resources to Iowa than to California if CC(cal) is more than (12 / 3.47) times CC(iowa), or about 3.46. I don't know if San Francisco and Los Angeles are three and a half times more expensive as media markets than Des Moines, but it's mathematically possible that Iowa would be the better bet.

Disclaimer: I tend to screw up off-the-cuff mathematics in some way, but the general point -- that under NPV a vote anywhere is worth just as much as a vote anywhere else, so it's the cost per net vote that should determine targeting decisions -- should hold in any event.

Posted by cminus | July 18, 2008 12:52 PM

Matt Weiner - thanks for the correction, it all makes much more sense now, and it resolves my concern about the discrete nature of electoral votes. It raises another issue, though. I had assumed that they meant that each state would divide its electoral vote in proportion to its own popular vote. Instead, it would seem, each state that signs on would (once the statute is triggered) allocate all of its electoral votes to the winner of the national popular vote (hence the name - and yes, I'm an idiot for not realizing this sooner).

Here's the problem - who decides what the national popular vote is? In a close election, is California really going to cast all of its electoral votes for the Republican based on the vote tallies of states like Florida? The same goes for Republican states casting their electoral votes based on the vote tally coming out of, say, Chicago.

Moreover, what do we make of different turnout rates in different states? There's an analogy here to the primary/caucus debate. Imagine hypothetically that Oregon's turnout is really high because of mail-in ballots. Then Oregon's vote will "count more" than a state with lower turnout, even though the lower turnout is the result of a legitimate procedural choice (it's not obvious that mail-in ballots are a good idea, particularly when used in a more interesting state like, say, Louisiana).

Posted by minderbender | July 18, 2008 1:28 PM

What worries me about this is what happens in extremely close elections; a recount could be a horrific mess, especially if the legal triggers for a recount in one state depended on the counts of votes in another state.

I also worry about extremely close elections. What happens if a state tries to game the system by pulling out (or joining) the NPV system at the last minute or even between the vote and when electors are determined?

Posted by Al | July 18, 2008 1:28 PM

The terms of the NPV compact, which would be enforceable in a court for the same reasons that the compact between New York and New Jersey creating the Port Authority is is enforceable (not for the weird contract clause reasons that the NPV people seem to think it is), prohibit pulling out after July 20 of a Presidential Election year.

Posted by washerdreyer | July 18, 2008 1:38 PM

The reason the "extremely close election" remark is kind of a canard is that its solution is the effective disfranchisement of large groups of people.

Imagine that a set of 1000 electors, chosen at random from the population, were the ones to decide who becomes President. We could recount a hundred times in a week, because those are the only votes that count. But only at the cost of disfranchising the entire US population.

But that's more or less what we do now. How is the current system preferable to one where we choose, in the event of a close election, a random one million votes, and recount those and only those to determine who wins? We make recounts more manageable only because no one except those in the single closest large state (OH, FL, whatever) end up counting.

BTW--in both 2000 and 2004, it was clear who actually won the popular vote. It is difficult to see how you would get to a few thousand nationwide. But if we do...is the best solution really just to disregard all those millions of votes?

Posted by Andrew | July 18, 2008 1:40 PM

But the NPV feels so good, you don't want to pull out.

Posted by Cake Giddy | July 18, 2008 1:47 PM

minderbender and Matt Weiner--I think you overstate the likelihood of an extremely close result in the national popular vote. The closest election in recent memory, 2000, featured a margin of over 500,000 votes between the winning and losing candidates. Even the 1960 election, which was ridiculously close, was decided by over 100,000 votes. In an election in which over 120,000,000 votes are cast, the possibility of a popular vote margin so close as to be conceivably affected by a recount (or rather, many separate recounts) is just vanishingly small.

As to minderbender's specific concern about whether states would recognize vote totals as reported by other states or localities, I think the answer is clearly "yes." Each state is already required to file an official report with the Archivist of the United States, certifying the popular vote totals from their states. I can't really conceive why you think it would be a problem--one state would just say "I don't know...I just don't trust those vote totals that came out of Chicago, as officially certified to the federal government" and would then go to court, and lose, to avoid recognizing them? Seems pretty far-fetched.

Posted by Anon21 | July 18, 2008 2:00 PM

I think you overstate the likelihood of an extremely close result in the national popular vote. The closest election in recent memory, 2000, featured a margin of over 500,000 votes between the winning and losing candidates.

Right. The NPV system would actually ENORMOUSLY REDUCE problems with fraud and with results that are "too close to call". That's because it's very hard to rig an election across 150 million voters and thousands of jurisdictions, and the chance that such an election will come down to a few hundred votes is beyond miniscule.

But elections in individual states might well come down to a few hundred or a few thousand votes. In the current electoral college system, where state electoral votes are winner-take-all, if you can swing those few thousand votes in a critical state you can pick up dozens of electoral votes and swing the entire election.

Posted by MQ | July 18, 2008 2:21 PM

Andrew-

Recounting the "single closest large state" only isn't a random choice. It's done (eg. Florida in 2000) when the other states electoral votes split evenly enough that the state in question is known to be decisive. Decisive, that is, according to the Electoral College, which is the current law.

If we chose one million random votes from a close election, there is no guarantee that they would be decisive, according to the method of choosing the plurality winner of the popular vote.

I do agree that the Electoral College is unfair and arbitrary, but I'm not sure how the fact that we need only look at a subsample of the entire vote once we know what's going on in the other states speaks to that.

Posted by EU | July 18, 2008 2:32 PM

What mechanism would actually determine what the national vote is. I suppose every state publishes its own popular vote and any idiot could add up the numbers. Still doesn't there need to be some authoritative agency to make this all official? For this to mean something a secretary of state in a state that voted for candidate A would have to certify electors for candidate B and whatever the law is they would have enormous political reasons not to do this.

Posted by Craig | July 18, 2008 3:58 PM

Craig: Again, the decision to certify the electors would not, in any meaningful sense, rest with the Secretary of State, any more than it does now. Upon the recognition of a winner of the national popular vote, the compact would come into effect, and the SoS would rubber-stamp the national winner as the recipient of that state's electoral votes. Any breach of that obligation could be remedied through legal action by other state parties to the compact. And frankly, it just isn't even slightly likely to happen. Do we worry, under the current system, that Colorado's Republican Secretary of State is going to give the big "fuck you" to his state's voters, should they choose Barack Obama, and certify the state's electors for his own preferred candidate, McCain? No, because political actors aren't just unconstrained by laws. They have to do certain things, even if they don't like them, because there's a law that compels them to. So it would be in the case of the National Popular Vote Compact.

As to the actual mechanism for certifying the winner, I can't see that as any kind of serious or complicated issue. As you say, anyone with a calculator could figure it out. Major news organizations will be reporting the same figures across the board, drawn directly from the numbers from each state's election officials. You are not going to see situations in which some states are arguing that one candidate won the popular vote, and others are arguing the opposite. It's an issue of fact--one candidate will be known to have won the popular vote and thus will receive the compacted electoral majority, and the other candidate will lose the popular vote and the Presidency.

Posted by Anon21 | July 18, 2008 4:30 PM

Take Utah--please. (Insert rimshot.)

No, seriously. You mentioned Utah, so take Utah as an example.

Suppose NPV had been in effect in Utah (and enough other states) in 2000. Utah would have ended-up having its electoral votes cast for Gore, even thoough 90% (or whatever) of its people are conservative Republicans.

I suspect that many (if not most) in that state would have preferred that their votes go for Bush as a matter of principle, regardless of what the National Popular Vote was.

Moreover, Utah's electoral votes (among those of other similar states) put Bush close enough to contest Florida and ultimately win the electoral college.

While NPV would have made things neater and cleaner in 2000, and the candidate I voted for would have won instead of Dubya, it would have caused states such as Utah to question their participation in NPV and may indeed have led to Utah withdrawing from NPV in future elections.

All hypotehtical, to be sure. But there are good reasons why some states wouldn't want to participate, so I doubt NPV will ever happen. And if it does, nothing will stop states from backing out if they become disgruntled with it.

Posted by Jim | July 18, 2008 5:06 PM

Jim,

There is no real way of knowing what the people of Utah would think in that scenario. Personally, I think most people tend not to care about issues such as what exactly happens with their state's electors. Rather, I think they more care who wins the election, and if they think the result was fair, I suspect that will be that.

And in any event, the chances of getting rid of the electoral college are negligible, as are the chances of getting every state to agree to simulating a NPV. So it is pretty much going to have to be something like this if the minority blocking reform is ever going to be defeated. In that sense, we might as well try.

Posted by DTM | July 18, 2008 5:20 PM

Utah is sort of a bad example. None of the sparsely-populated states with more electoral votes than their actual populations entitle them to would have any particular incentive to participate in this scheme. Inertia is the primary opponent of the NPV scheme, not small states as such.

I would agree with DTM that the concern for states opting out after election results they didn't like is overstated. I would expect the voters of Utah to be more angry at their fellow citizens for electing the wrong candidate than looking for some way to change the system in a manner that could conceivably disadvantage a candidate Utah voters favored in the future. Process issues are just hard to get voters worked up about, which is both the biggest obstacle to the NPV thing succeeding and the best guarantee that it would stick if it did succeed.

Posted by Anon21 | July 18, 2008 6:15 PM

each individual state would revert to having around as much weight as their population dictates. Which seems pretty fair.

Not quite true. It would have as much weight as the number of people from that state who vote in that election, which isn't the same. It's affected both by the difference between the population of the state and the number of eligible voters, and then how many of those voters cast ballots. So minors, non-naturalized aliens, convicted felons in some states, etc. aren't counted at all in the NPV system. Obviously they can't cast ballots in the current system, but they are counted in the total population of the state, and are thus weighted into their state's electoral vote.

Posted by Adam Villani | July 18, 2008 7:32 PM

The streamlined aspect of this plan is also its weakness.

Creating the system is very easy because it only requires states to pass ordinary legislation, a low procedural bar. But that low procedural bar works the other way too, with timely repeals constantly frustrating the goal.

Say you have states amounting to 300 electoral votes signed up, and one of them is Texas. What happens if a majority in the Texas legislature decides it doesn't like the candidate who's leading the polls in mid-August? You see where I'm going with this . . .

Posted by southpaw | July 18, 2008 7:45 PM

I don't like the NPV because of the potential for instability from one election to the next. Suppose that enough states enact NPV for it to take effect in 2016. But then one state's legislature flips in the 2018 midterms and switches back because the party in control would have an advantage under the old system. So we run the election as usual in 2020. But then the decennial reapportionment takes place in 2022, giving NPV states enough EVs to take effect. So the 2024 election is a popular vote election. But then the party in control of a large NPV state doesn't like the outcome of the election, so they revoke the compact after the election and send their own slate of electors, which is their constitutional prerogative.

Another issue is similar to one that popped up in the primaries. If we're going to count a total popular vote that's meaningful it should be the result of uniform rules across the country. In Wisconsin there are practically no barriers to voting; you can show up on election day with a piece of mail and a smile. In states like Indiana, however, you have to show ID and register months in advance. States also have vastly different residency requirements. Some states you can live there for a couple of weeks before the election and still vote, other states it's months. If we're going to have national elections, we need to have national standards, and the only way to do that is with a constitutional amendment.

Posted by Greg | July 18, 2008 8:08 PM

"Suppose NPV had been in effect in Utah (and enough other states) in 2000. Utah would have ended-up having its electoral votes cast for Gore, even thoough 90% (or whatever) of its people are conservative Republicans.

I suspect that many (if not most) in that state would have preferred that their votes go for Bush as a matter of principle, regardless of what the National Popular Vote was."

States get electoral votes for either their congressmen, or for their two Senators (the latter is why it disadvantages large states). You could create NPV so that the Senatorial votes are awarded to the national popular vote winner, while the House votes are awarded state by state, or vice versa. That is still enough for the system to work, since if someone wins the Electoral College and loses the popular vote, it will never be by a blowout in both places. That way Utah residents still see some of "their" electoral votes go to whoever carried their state.

Alternately, each state in the compact awards one of their electoral votes to the popular vote winner. The other still go to whoever carried the state. If this isn't enough for a candidate who gets a popular vote plurality to get an electoral college majority, each state then does the same with a second vote, and so on. The compact would only kick in if you had a different Electoral College winner from the popular vote winner, and then only to the extent that it would fix this. Most elections people wouldn't even notice it.

Note that in 2000 you would have needed just two red and two blue states to set things up this way, and it would have changed only two EVs, and that would have been enough.

Posted by Ed | July 19, 2008 6:12 AM

You guys are nuts. Does anybody worry that a popular vote is the wrong system for electing governors? US Senators?

Bottom line, the person who gets the most votes should. Furthermore, every voter should be equally meaningful. I don't want to call voters in Ohio. I don't want to drive to another state to campaign.
Every vote should count, every vote should be meaningful and the person with the most votes should win.

NPV accomplishes this, far as I can tell.

Posted by Chet | July 19, 2008 11:05 PM

You guys are nuts. Does anybody worry that a popular vote is the wrong system for electing governors? US Senators?

Bottom line, the person who gets the most votes should. Furthermore, every voter should be equally meaningful. I don't want to call voters in Ohio. I don't want to drive to another state to campaign.
Every vote should count, every vote should be meaningful and the person with the most votes should win.

NPV accomplishes this, far as I can tell.

Posted by Chet | July 19, 2008 11:07 PM