[EL] Two thoughts on the Electoral College and National Popular Vote
Sean Parnell
sean at impactpolicymanagement.com
Wed Nov 28 14:40:20 PST 2012
Jack - Again, we seem to have very different starting points and assumptions
on what "good" federalism and republicanism are, among other things.
The mathematical model you describe can be informative and interesting, but
it hardly proves anything, at least not to someone who favors the idea that
states as sovereign political entities should be responsible for selecting
the President. While it may be popular to say that "Ohio decided who would
be President after 2004," it is in fact wrong - Ohio, Florida, Texas,
Georgia, Iowa, Tennessee, Wyoming, Colorado, Missouri, etc. etc. decided who
would be President after 2004. Ohio was just one of the competitive states,
take away the 'safe' states and Ohio no longer 'decides' the election.
You assume that because you didn't see any television ads, Barack Obama
didn't want your vote and valued it less than he valued my vote, which he
did ask for. I do not share that assumption.
Best,
Sean Parnell
President
Impact Policy Management, LLC
6411 Caleb Court
Alexandria, VA 22315
571-289-1374 (c)
sean at impactpolicymanagement.com
From: Jack Cushman [mailto:jcushman at gmail.com]
Sent: Wednesday, November 28, 2012 3:58 PM
To: Sean Parnell
Cc: law-election at UCI.EDU
Subject: Re: [EL] Two thoughts on the Electoral College and National Popular
Vote
On Wed, Nov 28, 2012 at 10:03 AM, Sean Parnell
<sean at impactpolicymanagement.com> wrote:
As to my vote being 44 times more "influential" - again, I do not agree with
that. My single vote goes towards allocating Virginia's 13 electoral votes.
Your single vote goes towards selecting Massachusetts' 11 electoral votes.
The most significant difference between us is that I must suffer through far
more mail, radio, and television than you. For what it's worth though,
According to FairVote in 2004 and 2008 your vote in Massachusetts was
actually of greater value than mine in Virginia and Illinois. See
<http://www.fairvote.org/assets/Uploads/npv/2008votersperelector.pdf>
http://www.fairvote.org/assets/Uploads/npv/2008votersperelector.pdf. Didn't
find similar info for 2012, but someone on the list may have it.
I should apologize -- I haven't spelled out the basis of my claims about
voting power, and they're key to understanding my problem with the Electoral
College. When I say "44 times more influential," I'm citing the result of
the best model I'm aware of comparing your voting power to mine in the 2012
election. Can we set aside our swords for a while and figure out this piece?
If the math is wrong, I'd love to know, but I think it's pretty solid, and
the conversation doesn't make any sense without it.
So let's talk voting power. From an intuitive standpoint, it's been
popularly understood at least since the term "swing states" exploded in 2000
<http://books.google.com/ngrams/graph?content=swing+states&year_start=1800&y
ear_end=2008&corpus=15&smoothing=3&share=> that candidates themselves do
not value all votes equally. As a candidate, I need to spend my limited time
and money wisely, and if changing some minds helps me more than others I'll
focus on those people. This isn't a subtle effect -- in the last four
elections, the campaigns have spent nearly all of their time and money in
about a dozen states <http://www.fairvote.org/presidential-tracker> .
They're not dumb people. If my vote was of greater value than yours (or of
any meaningful value) they'd make some effort to influence me. But the
states they focus literally all of their efforts on have essentially no
relationship to that PDF you linked. So how are they deciding how to
allocate their resources? They're using internal computer models, based on
polling data, that assign a relative worth to the voters in each state.
This is "voting power" -- the relative likelihood that any given voter will
decide the outcome of the election. In a simple democratic election, every
voter has identical voting power of 1. In a tiered election, the tiers
reallocate voting power from some voters to others, and things get really
tricky to quantify. As explained in this 2002 paper
<http://www.stat.columbia.edu/~gelman/research/published/STS027.pdf> , the
easiest mathematical models assume that each voter flips a coin to decide
who to vote for, and that yields statistical results that have little
relationship to reality. (Coin-flip models predict that large states will
have more power; your votes-per-elector model might predict that smaller
states will have more power. Neither of those is borne out by election
results.)
In order to get a useful model of voter power -- a model that, if you were a
candidate, you'd want to put money on -- you need to run thousands of
simulations based on actual public opinion in the current election,
perturbing the results based on a realistic model of how the polls might be
wrong and how blocs of voters might shift their opinions. The results of
your simulation will tell you which polls you should care about and which
ones you shouldn't -- in what states will the smallest shift in your polls
have the largest impact on your odds of success?
Until recently, I don't know that there have been good public models of
voting power for the US presidential election. But it turns out the 538 blog
uses a simulation like the one I described above to predict electoral
outcomes, and in the last election they published a measure of voting power
under the heading "Return on Investment Index
<http://fivethirtyeight.blogs.nytimes.com/> ." (You have to scroll about
halfway down the page, on the right.)
Their model clearly explains why campaigns are targeting some states so
decisively. If the average voter has voting power of 1, we might expect
about 25 states to have power a little above 1, and 25 states to have power
a little below 1, right? But only eight states have power of 1 or more. 38
out of 50 states have power of 0.1 or less. That means if a candidate wins
one voter in any one of the top five states, while losing one voter in all
40 of the bottom states, they have increased their chances of winning the
election.
Virginia has voting power of 4.4; Massachusetts has voting power of less
than 0.1. So when I say you have 44 times as much influence as I do, here's
what I mean: either of the candidates in the 2012 election would have
happily lost my vote and the votes of 43 of my neighbors if it meant
changing your mind the other way. If we both volunteered to canvass, I could
persuade every voter on my street, you could persuade your next-door
neighbor, and you would be the more valued member of their campaign. Your
involvement in the civic process where you live is relevant and encouraged
in a way that mine is not.
This is an oversimplification, because those numbers won't stack additively
exactly, and the underlying model isn't perfect. But it's the best available
measurement I know of, borne out by the way the campaigns actually behave.
We'll come closer to the truth if we assume these numbers are right-ish, and
decide whether that's a problem, than if we assume anything else.
So to pick up my advocacy sword again. Tara observed:
Had Cleveland won that year, he would have done so because he ran up his
vote totals in the South. Why should a handful of Southern states be able
to pick a President for the rest of the country? Why is that purely
democratic outcome more fair?
The question is, in that scenario, how would Cleveland have run up his vote
totals? He would (in modern terms) have adopted positions favorable to
southern voters and spent more of his money in the South -- but he could
only do so if his positions did not lose him an equal number of voters in
the rest of the country, and the South remained the most cost-effective
place to spread his message. He would be pulled in a tug-of-war between the
persuadable voters in the South, and the persuadable voters in every other
state of the Union. It's fair because losing my vote in Massachusetts would
matter precisely as much to him as picking up another vote in South
Carolina, and he'll be my president precisely as much as theirs.
Compare that to the way our candidates are insulated from national opinion.
When Obama turned in his first debate performance, he was less worried than
his supporters, because he knew something they didn't. He knew which voters
were persuadable in Ohio, he knew which of those were changing their poll
answers after the debate, and he knew he wasn't losing ground in the key
persuadable voting blocs that would secure his election. (Google isn't
spitting out the cite for this, but I know one of those post-election
articles dug into it.) In a nation of 300 million, he cared about the
opinions of maybe 300,000 -- or 1 in 1000. How did people in Texas respond
to the debate? Didn't care, didn't ask.
Which brings me to:
Jack: suffice it to say we see things very differently in regards to
republicanism, federalism, the role of the states, and the electoral
college.
I'm really not sure what it would mean to say this system promotes
republicanism. Does it serve in some way, for example, the principle that
our elected representatives are sometimes wiser than ourselves, or that the
majority of citizens cannot deny the fundamental rights of the minority?
It's not just different viewpoints -- I'm actually not understanding what
that word means to you.
I do see that it promotes federalism, of a kind: it grants extraordinary
importance to the lines between Florida or Ohio and neighboring states. But
I don't see how it promotes a good federalism. Did we mean, by federalism,
that the politically active citizens of 38 states would ignore their fellow
state citizens, and instead travel to a few chosen states that control the
outcome? It's stirring to hear stories of volunteers from around the country
who dropped everything to staff a field office in Ohio, because that was the
closest place they could help -- but is that the federation we wanted to
create? Some states you come to for money, other states you come to for
votes?
To the contrary, I think the NPV-IC is federalism working. If a majority of
citizens in my state say, "this swing state phenomenon isn't how I saw our
federation going; under my vision of fairness, the citizens of all the
states should have a more or less equal voice in the presidential election,
and I direct you to assign our electors to bring that about," we are using
the power granted to us as a state to adjust for a dysfunction in the
federal Constitution's electoral process. If we try that and don't like how
that assignment goes, we can direct our state in the next election to assign
our electors by House districts, by the legislature's whim or by trial by
combat. They're our electors.
I mean, states' rights are not a universal balm; it might be wise at some
point to shift the presidential-election power from the states to the
federal government by constitutional amendment, as we have time and again
shifted power over voting processes to comply with an evolving concept of
fairness. But what we're talking about with the Compact is state authority
being used to promote the state's interests -- anti-federalist it's not.
Best,
Jack
As for vote trading - people are generally free to vote based on whatever
factors they chose, for good or ill (with certain exceptions - paying people
to vote being the most obvious). I haven't given it much thought though, and
so don't have much of a response to you other than to say lots of people
think their vote "doesn't count" - ask any liberal Democrat in Northwest
Iowa or Republican in Chicago how much they think they're vote "counts" for
Congress or Mayor, respectively, and you're likely to get answers similar to
those folks complaining that their Presidential vote in Utah or Vermont
doesn't "count" because candidates haven't asked for it.
Best,
Sean Parnell
President
Impact Policy Management, LLC
6411 Caleb Court
Alexandria, VA 22315
571-289-1374 (c)
sean at impactpolicymanagement.com
From: Jack Cushman [mailto:jcushman at gmail.com]
Sent: Tuesday, November 27, 2012 6:18 PM
To: Sean Parnell
Cc: law-election at UCI.EDU
Subject: Re: [EL] Two thoughts on the Electoral College and National Popular
Vote
On Tue, Nov 27, 2012 at 3:12 PM, Sean Parnell
<sean at impactpolicymanagement.com> wrote:
Jack - I like my democracy tempered with a bit of republicanism, such as the
Bill of Rights, U.S. Senate, separation of powers, etc.
More importantly I've always found it silly to say that the votes of people
in Massachusetts or Texas don't count when it comes to presidential
elections. Of course they do, just like the votes of people in reliably
"safe" Congressional districts count when it comes to determining who will
be elected Speaker of the House. Lack of competitiveness in any particular
state or district is hardly indicative that democracy has failed, or that
the principle of equal voting rights is being trampled.
Best,
Sean Parnell
There are two different systems I can understand here for selecting a
unitary official like a president, and I'm not objecting to either of them:
-- In what I'll call a republican election, citizens select representatives
(each citizen having one equal vote) to choose a leader on their behalf;
their representatives choose the best Speaker of the House, Prime Minister,
etc. It's a useful system when voters lack the knowledge to select the
ultimate leader, or when the best choice requires the kind of negotiation
that only small groups are good at, or when (as in the Senate) you set out
to privilege some voters over others.
-- In a "democratic" election, citizens choose a leader by equal vote
directly.
I get those, I think. They're both useful ways to make decisions. But the
Electoral College is neither. Voters aren't choosing local representatives
to make a decision on their behalf, they're directly choosing the unitary
official to be elected -- but their preferences are grouped geographically,
so that changing the minds of voters in some regions is worth more than
changing the minds of voters in other regions. It's still fundamentally a
democratic-style election, delivering none of the benefits of a republican
election I can think of, but it's modified to randomly make some voters more
important than others.
No one set out to do that; no one designed the system to privilege voters of
swing states in a national presidential election, the way they intended to
privilege voters of small states in the Senate. It's an unintended
consequence of democratic selection of electors, and of the power of
faction, and I suppose of national polling.
And this is where you lose me: if this is an accident of history, why is it
a happy accident? From the standpoint of political ideals, not
administrative convenience -- what principles of federalism or republicanism
are served by a system where we both choose from the same slate, but your
preference for a President is 44 times more likely to influence the outcome
than mine? If we're all choosing from the same list of names, what
republican goal is served when candidates will do anything for votes in
Virginia and Ohio, while making no attempt to campaign in 38 of the states
they are to lead?
Maybe it would help me to understand this if you could point out some other
democratic election run the same way, where citizens choose from the same
list of ultimate outcomes, but some voters have more influence over the
outcome than others. This forum (and this thread!) constantly remind me how
much I don't know -- are there other elections where we happily accept this
kind of imbalance?
As another angle, which may or may not be helpful -- if you (for example)
were a Republican in Ohio, would you trade presidential votes with a
Democrat in Massachusetts? You have to stand in line to press the button for
Barack Obama, they have to suck it up and connect the arrow for Mitt Romney.
What would it take to get you to do that? Conversely, if you were a
Republican in Texas, would you trade votes with a Democrat in Virginia?
Would you feel differently about the two trades? And if so, is there a
reason we benefit from having 1/4 of the country think of their vote one
way, and 3/4 of the country think of their vote the other way?
Best,
Jack
Best,
Sean Parnell
President
Impact Policy Management, LLC
6411 Caleb Court
Alexandria, VA 22315
571-289-1374 (c)
sean at impactpolicymanagement.com
From: Jack Cushman [mailto:jcushman at gmail.com]
Sent: Tuesday, November 27, 2012 2:24 PM
To: Sean Parnell
Cc: law-election at UCI.EDU
Subject: Re: [EL] Two thoughts on the Electoral College and National Popular
Vote
On Tue, Nov 27, 2012 at 11:51 AM, Sean Parnell
<sean at impactpolicymanagement.com> wrote:
1. Apparently only 17 states have completed their count of all
ballots, per this USA Today editorial (as a rule, I abhor citing editorials,
but I'm going to trust they got this fact right):
http://usatoday.com/story/opinion/2012/11/26/counting-votes-voting-system/17
28529/ I think the implications for National Popular Vote are pretty obvious
- had this been a closer election (say, Bush-Gore or Kennedy-Nixon close)
we'd still not know who the president was, and there would be horrific legal
battles being waged right now all across the country over which ballots
should or should not be counted. The Electoral College seems to have
provided conclusive clarity rather quickly.
But the states this year have no particular reason to hurry in certifying
their results. And in 2000 the election wasn't decided until December 12,
over two weeks from now. I don't see why we couldn't resolve the legal
challenges, run the recounts, and certify an official national popular vote
in a close election at least as quickly as the Electoral College was decided
in 2000. These are problems that are resolved in parallel, not consecutively
-- why should other states take longer than Florida did?
The certified totals might ultimately prove to be incorrect, as they were in
Florida in 2000. But as 2000 shows
<http://en.wikipedia.org/wiki/United_States_presidential_election,_2000#Post
_recount> , the Electoral College can't save us from simply certifying the
wrong winner. If we don't like slow, uncertain, inaccurate elections, the
solution is to adopt national standards for modern, reliable voting
processes. We're a wealthy technological nation, and it's an eminently
solvable problem.
But set that aside for a minute. The more important point is that it's
better to have horrific legal battles and democracy than no horrific legal
battles and no democracy. Because you know what's easier to run than a
democracy? Any other system of government. "Easy" isn't what democracy is
about.
Counting every vote is a difficult logistical problem. The Electoral College
eases the vote counting process by ensuring that 3/4 of states -- and their
voters -- are essentially irrelevant to the election; care and attention can
then be focused on the minority of voters with the lion's share of voting
power. That's the core of this kind of practical objection to the National
Popular Vote: it's too much of a hassle to accurately count my vote in
Massachusetts or Tara's vote in Texas, so we should avoid that necessity by
giving Sean 44 times <http://fivethirtyeight.blogs.nytimes.com/> as
important a vote in Virginia, and focus on counting his.
But that's not who we are, and it's not what the Electoral College is for.
It began as a means for state legislatures to select representatives to
debate and choose a President -- a republican strand of our democratic
republic. It lingers, zombie-like, to capriciously reallocate voting power
from some voters to others in a popular election. It's not a structural
choice; there is no rhyme or reason to the states it chooses to favor. It
gives us the questionable benefit of ignoring the shoddiness of the voting
systems in many parts of our country, but it cuts against a principle we
have consistently sacrificed for: when a group of citizens select one among
us to be our leader, each of us is entitled to an equal vote.
It would be great (for so many reasons) if we could first build an
effective, reliable voting system, and then adopt an equal vote. But that's
never been how things go. First you win the right to vote, and only then --
if then -- do they build the voting booths. Better long lines than no lines
at all.
Best,
Jack
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20121128/2f7690e5/attachment.html>
View list directory