Electoral College Bias
It could all come down to Pennsylvania.
That US presidential elections are decided through the mechanism of the Electoral College which, with two minor exceptions, awards all of the Electors of each of the 50 states (the number of House seats, which varies by population, and Senate seats, which is two per) plus the District of Columbia (fixed at three per the 23rd Amendment) to the winner of that state’s votes, regardless of margin, will not be news to OTB readers. Not only are you reading a political blog, and thus likely to have known this for a very long time, but it’s a topic we’ve harped on repeatedly for the last two decades.
In “The systemic bias Kamala Harris must overcome in order to win,” The Economist breaks down just how much this matters.
This electoral system is an obstacle for Ms Harris, the Democratic nominee, as it was for Joe Biden and Hillary Clinton before her. Although our presidential forecast gives her a 71% chance of winning more votes than Mr Trump nationwide, she has only a 51% chance of winning the electoral college.
The electoral college is formed of electors from every state and the District of Columbia, roughly in proportion to their population. To get to the White House, a presidential candidate must win a majority—270—of the 538 electors. Aside from those in Maine and Nebraska, practically every elector votes for the winner of the popular vote in their state. This winner-takes-all system encourages candidates to focus on swing states, rather than on places where the outcome is pretty much assured.
This year, we estimate that Pennsylvania is the most likely state to be the “tipping-point”: it provides the 270th elector for Mr Trump or Ms Harris in 26% of forecast simulations. Pennsylvania is followed by Michigan, North Carolina and Wisconsin as the likeliest pivotal states. On average, the tipping-point states are whiter and poorer than the country as a whole. All this shapes candidates’ stances. Ms Harris has recanted her opposition to fracking, a big employer in Pennsylvania, for example. Both candidates have endorsed eliminating taxes on tips, which could play well with hospitality workers in Nevada, another potential tipping-point state.
We’ve seen the distortions of the peculiarities of our electoral system in myriad ways. Notably, the fact that Iowa led off both the Republican and Democratic primaries for half a century (until Biden’s team changed it for the irrelevant 2024 cycle) meant a whole lot of pandering to Big Corn. But that Harris has changed a signature policy because of the outsized power of Pennsylvania voters and that Trump came up with one of the dumbest ideas ever (and Harris jumped on the bandwagon!) to suck up to Las Vegas has drawn less attention.
Regardless, the main issue is that, for all intents and purposes, the so-called “swing states” choose our presidents rather than the American people writ large, which is the most notable distortion.
Though Ms Harris leads in our tracker of national (pre-debate) polls by 2.5 percentage points, her margins are narrower in swing states.
This should not shock anyone. The thin margins are what make them “swing states,” after all.
In Pennsylvania she leads by 0.7 points, according to FiveThirtyEight, an aggregator. By itself, this suggests the electoral college has a bias of 1.8 points—smaller than the 2.9 points in 2016 or 3.3 points in 2020.
That’s an odd way of describing it. The bias created by the Electoral College shouldn’t be impacted by the closeness of the polls in a given state. While it’s technically true that, if Harris won every state, the Electoral College outcome and the national outcome would be the same, the distortion is to make small-population states more important than they ought to be (because each state gets two Senate-driven electors regardless of population, whether it’s Wyoming’s 584,000 or California’s 39 million) and makes voters in “safe” states irrelevant (because of the wnner-take-all distribution).
So, for example, in 2020, Biden received 3,458,229 votes in Pennsylvania to Trump’s 3,377,674 (50.01% to 48.84%)—a difference of 80,555 votes (1.17%). In 2016, Trump received 2,970,733 votes to Clinton’s 2,926,441 (48.18% to 47.46%)—a difference of 44,292 (0.72%). In both instances, the winner got 20 Pennsylvania Electors and the loser 0.
By contrast, Biden won California 11,110,250 to 6,006,429 (63.48% to 34.32%)–a whopping 5,103,821 (29.16%) margin. Clinton won 8,753,788 to 4,483,8810 (61.73% to 31.62%)—damn near doubling Trump’s vote and percentage. In both instances, the Democrat got 55 Electors and Trump 0. But here’s the thing: far, far more people voted for Trump in California than in Pennsylvania both times. Their votes are completely discounted under our system.
Alas, the rest of the article isn’t particularly helpful, as it doesn’t really explain the model.
But state opinion polls should be taken with a pinch of salt. In 2016 they indicated that Mrs Clinton was on course for victory. Although polls correctly suggested she had a small lead nationwide, they systematically overestimated her strength in swing states. Mr Trump won despite getting 2.9m fewer votes in total. Four years later, after months of favourable polling, Mr Biden won by the narrowest of margins in Wisconsin and Pennsylvania.
To hedge against this, our forecast combines polling with a “fundamentals” estimate of how states will lean, relative to the country, for example taking into account their tilt at the last election. The model believes Ms Harris will face an uphill struggle in the electoral college similar to Mr Biden: she would need a 2.1-point lead in the two-party popular vote to be the favourite to win the electoral college (see chart).
I’m not sure how helpful said chart is:
The standalone model page shows the overall and state-by-state but no how it’s derived. But, as of this morning, is shows Harris with just enough Electors to win (272) but with a range of anywhere from 191 to 377. Trump has 266 Electors and a range of 161 to 347. You’d rather be in Harris’ shoes than his but it’s still essentially a toss-up.
And, once again, while we have 50 states and the District of Columbia participating in the election, the Economist concurs with pretty much everyone that only seven states are in play. Here’s their model’s current assessment:
Harris is certainly doing better than Biden was, as Georgia and North Carolina were rather solidly in Trump’s column before Biden exited the race. Still, it’s remarkable that the contest is even in doubt given Trump’s performance in office, his attempt to steal the last election, his felony conviction, and all the rest. But here we are.
The odds that the former guy “should be” elected are (IMHO) zero, no matter how many Californians cast votes for him.
Typical gerrymandering involves “crack and pack,” which involves creating “wasted votes” by creating districts that lean heavily one way or the other, such as urban districts that vote 90% D, or rural 80% R districts..
It’s true that small states get disproportionate EC votes, but if, as it may happen, they vote heavily towards the same one party, as they currently do, they also contain a lot of “wasted votes.”
Which party benefits from EC bias is a transitory phenomenon that can vary from time to time.
What happens, for example to the EC, if TX tilts blue, which I think is coming sooner than many people think.
@charontwo: My problem isn’t that it’s biased to a particular party but against democracy. Before it flipped, seemingly permanently, in 1992, California was a reliably Republican state. Indeed, Republicans were said to have a “lock” on the Electoral College because there were so many safe Republican seats that it was difficult for Democrats to find a path to victory.
I’m not sure that, as of this morning, it actually is that close. Yesterday, Ann Selzer released an updated poll in Iowa. Trump led by 4 pts. He had been leading Biden by 18 but, more informatively, in the final 2020 poll, she had Trump +7 (Trump won by 8).
If Trump is only +4 in lily-white Iowa by what has long been recognized as the “gold standard” pollster, than WI/MI/PA should all be a lot more secure than the Economist is projecting.
I might also note that reviewing the recent national polls on 538 the Economist is a significant outlier – possibly because their poll was pre-debate. They have the national race tied. Almost everyone post-debate was around Harris +4 (with more +5 than +3). The sole exception, Atlas, is just a weird result (Trump +3).
I don’t know if it would take a constitutional amendment, but one thing that could be done is a Federal law mandating that a state’s electoral vote be divided in proportion to the percentage of the popular vote statewide. It would make it worthwhile for a Republican to campaign in California and New York, and for a Democrat in Texas and Florida.
I think this is somewhat of a misconception about what creates the imbalance between the EC and popular vote. As I’ve explained before, in 2004 and 2012, the EC was slightly skewed in the Dems’ favor. So, contrary to popular belief, the pro-GOP skew in the EC isn’t something consistently true in the 21st century so far; it’s just something that’s been more noticeable due to there being two elections in this century in which a Republican won the EC and lost the PV, and no elections the other way around.
The main impact of the GOP’s domination of smaller, rural states is in the Senate. What causes an EC skew has more to do with which states happen to be the swing states in any particular election. For instance, Biden won WI in 2020 by about 20,000 votes. If we took that 20K out of Wisconsin and redistributed it to Maryland–a state roughly equal in population–he’d have lost WI and received no electoral benefit from the 20K, since Maryland was going to vote for him anyway.
Broadly speaking, we say 2020 showed a pro-GOP skew because the margin of Biden’s national victory (4.5 points) was a lot wider than his margins in the states that put him over the top in the EC. In contrast, Obama in 2012 had a slightly smaller national lead (3.9 points) but for Romney to have won the EC, he’d have needed at least one state that Obama won by over 5 points.
Therefore, any partisan skew in the EC is a lot more ephemeral than is commonly believed. Since there seems to be less of a difference between national polls and swing-state polls this time, that suggests the pro-GOP skew is smaller than it was in 2016 and 2020. The polls could be wrong of course, but the thing is, the skew was noticeable in 2016 and 2020 polls before the elections happened, so it seems that whatever drawbacks polls have, they do accurately pick up this effect.
The EC is a hangover from the undemocratic Constitutional Convention and concern for state sovereignty. It no longer serves a purpose. There is the National Popular Vote Interstate Compact which could de facto eliminate the EC. But that’s not going to happen as long as GOPs depend on the EC to have a chance.
The only way the NPVIC is going to be sufficiently adopted in the wake of a GOP winning the popular vote and losing the EC. I wonder if we could engineer that with an initiative to make only people with even SSNs eligible to vote in CA.
After more careful thought, I’m not sure I am against making tip wages tax-free. There are several arguments in favor that I think carry some weight:
1. Tip reporting is notoriously incomplete and inconsistent, precisely because reporting that income costs you money. Enforcement is difficult and expensive. If tips were tax-free, reporting (and thus information used to drive policy) might be more accurate. Employers would also owe matching payroll taxes on that additional reporting at the margin, increasing Social Security and Medicare revenues.
2. A tax break on tips earned is clearly progressive. Affluent people do not receive tips. The stimulus value of a tax cut on tips is about as high as you can get.
3. Total reported tips are around $30B per year. The personal income taxes collected on that are not really visible in the pie chart of federal tax revenues. If elimination of tax on tips came with a quid pro quo increasing taxes slightly on the wealthy, gross revenues would surely increase.
Of course, Congress* would need to firm up their definitions of what counts as a ‘tip’ to avoid shenanigans in contracting. “I paid $300k for my house, plus $200k in tips…”
*I started to type “the IRS would…” but then realized that, thanks to the Suborned Court, the IRS no longer has the authority to make such definitions binding on the courts.
@Kylopod:
Our current near parity of the parties should also be ephemeral. It’s an historical oddity. Normally our two party system has produced a nationally dominant party and an opposition party, sometimes described as a sun party and a moon party. From 1932 to around 1970 Dems were the sun. The national popular vote in 2020 was just about 51-47. In 1940 the prez vote was 55-45, nobody cared that the EC was a hugely distorted 85-15.
Harris is “here in Pennsylvania” but her demeanor says it’s a chore imposed by that damn constitutional majority empowered by the Electoral College system. She’s just not a woman of the people.
@gVOR10:
I agree. I think the closest era to the present is probably the late 19th century when elections were very close nationally, and the main reason the GOP was able to dominate electorally was due to the string of state-level gerrymanders where they got a bunch of low-population territories in the west ratified as states (including separating Dakota into two states for no good purpose other than as a raw power grab). The Dems for their part engaged in the massive disenfranchisement project in Jim Crow. Obviously not an exact analogue to today given the partisan swap over race, but that was the longest period of close elections until the present day.
@SC_Birdflyte: Are you aware of the National Popular Vote Compact? Just need a few more states to pass it (although little progress is being made …)
@DrDaveT:
They will now that the Leo & Crow court decreed bribes are legal if given as gratuities.
James wrote:
Leaving aside Trump’s claim that he only lost CA in 2020 due to election fraud, the substance of this observation is spot on. The same observation can be made for the 3,251,997 of my neighbors here in New York State who also voted for Trump. None of their votes ultimately “counted” which is ironic given how so many of those votes came from the type of rural “Real America(TM)” areas I have been reliably told that the electoral college is supposed to help support.
@JKB:
So all you can bring to the discussion is… vibes? Man, the Democrats are doing better than I expected.
As I will once again suggest, given your past opinions about women, maybe you should sit this one out. I have yet to see anything that suggests that you have a credible read on what most average people are looking for from Harris.
If it helps, based on polling, I think you’re a better bellwether on Republican men and white men who haven’t gone to college. So that’s something.
The Electoral College system drives me nuts specifically because of winner take all by state. That is a system problem. If your preferred candidate lost your state by one vote your vote flat-out doesn’t count. Systemic disenfranchisement.
Many states are flat-out gimme states. If you are a minority voter in your highly R or D leaning state, your vote just doesn’t count at all. Why bother? Even in a toss-up state, if your candidate loses, it as if you didn’t vote at all.
The EC system is, in my mind, voter disenfranchisement. It’s a system problem. And has a lot of unintended, negative consequences.
I know it would take a constitutional amendment to change, but the current system is broken. When you know your vote is meaningless, why vote?
Why not a national plebiscite like a rational nation?
If you are a D voter in Wyoming, or an R voter in Massachusetts actually showing up to vote for your preferred Presidential candidate is basically a chump move. There is no point.
@de stijl: I really think you’d be disappointed by a proportional electoral college. Think Red States don’t change much while Blue States with split legislatures–such as Oregon, Washington, and California (?) give 1/3-1/2 of the electoral vote to the Republican candidate. I’ll agree that a straight majority vote would be a superior system, but we’re not going to get there during my–or your–lifetime*, and the national vote compact scheme will be found unconstitutional by the current Supremes.
*I would add “or our grandchildren’s’ lifetimes,” too, but from what I understand, neither of us have any.
The reason that Iowa and New Hampshire have a greatly oversized role in candidate selection is that someone has to go first. They just happen to be / arranged themselves to be the first.
In my mind, the role of the first in the nation is to eliminate the never gonna happen folks from the future candidate pool. Sorry, ain’t gonna happen. (Which is an extremely useful function.)
As an Iowan from 2003 on I can verily confirm that being first in the nation absolutely sucks as far as TV ads and increasingly in texts.
The run-up before they held the Iowa caucuses I got probably several thousand spam solicitation texts. TV is nigh unwatchable unless you have a much greater tolerance than me. Mute button during ad breaks is essential It is so very annoying. Dozens of texts per day and it ramped up crazy hard as the day drew near.
Being first in the nation on the path to candidate selection really fucking sucks. Seriously, you all do not comprehend how overwhelming and annoying it is.
The dumbest thing is that so slap-dash on texts – I don’t believe I have ever voted for a Republican in any Iowan election, but I only got R texts. I’m the wrong audience.
@Just nutha ignint cracker:
I just want a flat-out national plebiscite. Take what state you live in out of the vote counting system.
At one point women voting required a national re-recokning. Former slaves. We could do it again. It’d be difficult. We should.
The EC system will haunt us forever. It’s so fucking stupid and undemocratic.
You know, just a straight up vote, like a normal country.
@JKB: Ah yes, the bullionaire who got rich thanks to daddy, went to an Ivy League school and has consistently screwed over everyone who worked for him is definitely the working man’s guy and definitely not the guy who will screw them over again if he gets a second term the same way he did in the first term.
Trump is the politician for trust fund babies and used car dealers. That’s why Kid Rock loves him – he’s the heir to an auto dealer fortune.
Just a friendly reminder that there’s nothing sacrosanct about the electoral college. It was one of the last cobbled together compromises made by the constitutional convention when everyone there was hungover, nobody agreed on how to elect a president but also nobody really cared because they all knew that no matter what the mechanism, Washington would be the guy and they were fine with that.
California has more Republican voters than Texas. Texas has more Democrat voters than NY. NY has more Republican voters than Florida. Florida has more Democrat voters than…I don’t remember which state was next in the parade.
The EC simply never envisioned a country with such wildly different populations by state. A rational country would fix that (which is why we won’t). In the end, size matters 😉