Burning the Midnight Oil for Progressive Populism
Lots of puzzling around about Romney tapping Ryan for Vice President. Lots of coverage of nervous Republicans down-ballot: People's World: Republicans fear Ryan pick could sink GOP.
Now, for inside baseball politics at the national level, I get nothing that can't be picked up by following the right people on twitter. However, I was thinking ... is it likely Romney picked Ryan without his campaign running the numbers? Maybe is there's something in the numbers that led them to pick Ryan, then I've got a shot of seeing tha footprints of that for myself.
When I want horse race numbers, I go to Nate Silver's FiveThirtyEight, taken into the NYT system last year. Now, I wouldn't go so far as to subsidize the unthinking stenography on economics and energy and transport that we normally think of when we think "New York Times" to get access to horse race info, but luckily the main blog is a teaser sitting outside the NYT paywall.
So, what tale do the numbers tell? Join me, below the break.
Swing Voter Swing States versus Base Turnout Swing States
Nate at fivethirtyeight has looked at which swing states are "swing voter" states and which are "base turnout" swing states ~ that is, they are swing states despite having few persuadable voters, because the bases are roughly equal in size.
In this article, Nate Silver develops an index he calls "voter elasticity", which is based on individual voter level exit survey data. He uses that to determine whether individual voters in a given state are more or less likely to "swing" than the national average after accounting for measurable demographic factors that correlate with the vote.
Now, a simplistic view of a swing state is a state that has fewer than 50% Democratic and Republican affiliated voters with swing voters in the middle either in the habit of splitting their vote or willing to vote for one party in one election and the other party in another. And looking at the voter elasticity, some states that are swing states have a relatively high voter elasticity. I'm calling those "Swing Voter" states. In terms of the traditional Presidential Politics "swing states" that Nate Silver picks out in his article, the "swing voter" swing states are:
- Colorado, Iowa, New Hampshire, New Mexico and Wisconsin
Not all high voter elasticity states are swing states: a state can have a high voter elasticity and a high partisan bias, so that it is normally a "safe" state, but every once in a while has an surprising result that goes against the normal outcome.
As noted above, there is a second way to be a swing state. One can have fairly few persuadable voters, and still be a swing state, if the electoral bases in the two states are fairly evenly balanced, so that statewide election results are determined by relative turnout. I'm calling these the "base turnout" swing states, and the "base turnout" swing states are:
- North Carolina, Virginia and Pennsylvania
Now, elasticity is a continuum, and given that there are some states that are more elastic than the national average and others that are less, there will be some that are close to the natural average. Roughly speaking, these states can be thought of as swinging by some combination of persuadable swing voters and differences in base turnout, so I am going to call these "balanced" swing states, which are:
- Florida, Michigan, Minnesota, Nevada and Ohio
This distinction between swing voter and base turnout swing states is the concept that I am bringing into this analysis.
The 2012 Underlying Political Terrain
All the states not mentioned are "lean" and "strong" states for each side: if one side is winning lean states from the other side, then its not a close election, but is moving toward landslide territory, and if they pick up any base states, then it is a landslide:
- The Republican Base for 140 electoral votes is Alaska, Arkansas, Alabama, Idaho, Kansas, Kentucky, Louisiana, Mississippi, Nebraska, North Dakota, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Utah, West Virginia and Wyoming.
- The Lean GOP states, adding 51 votes for 191 electoral votes, are Arizona, Indiana, Georgia, Missouri, and Montana
- The Democratic Base is 149 electoral votes: California, Connecticut, DC, Delaware, Hawaii, Illinois, Massachusetts, Maryland, New York, Rhode Island and Vermont
- The "Lean Dem" states, adding 37 for 186 electoral votes are Maine, New Jersey, Oregon, and Washington
I've used the site 270towin to illustrate the Base versus Base electoral map, GOP 140 vs Dem 149. The current electoral terrain is, however, slightly tilted to the GOP with Base+Lean favoring the GOP 191 to 186. However, because of Romney's "Latino problem", NM is not considered in play at the Presidential level, so taking it off the swing state list, that set of swing states puts Obama at 191 and Romney at 191, each needing 79.
So, now I line up the states in order of Nate Silver's current projected likelihood of the incumbent President winning the state. Note that this is not the project margin, but the likelihood that the margin is 50%+1.
Lining those up in terms of the 538's model of the likelihood of a win of that state in November,
Obama. I present it as a "incumbent path of victory", electoral votes at that step of the ladder as a fraction of electoral votes required, but obviously Romney's path to victory entails pushing the incumbent number below 270:
201/270 MN: 91.2% +10 ~ balanced
217/270 MI: 89.0% +16 ~ balanced
237/270 PA: 86.2% +20 ~ base turnout
247/270 WI: 80.1% +10 ~ swing voter
253/270 NV: 78.2% +6 ~ balanced
257/270 NH: 73.5% +4 ~ swing voter
275/270 OH: 67.9% +18 ~ balanced
288/270 VA: 66.2% +13 ~ base turnout
294/270 IA: 63.6% +6 ~ swing voter
303/270 CO: 59.2% +9 ~ swing voter
332/270 FL: 53.9% +29 ~ balanced
347/270 NC: 32.2% +15 ~ base turnout
Evidently the path to victory on that ladder runs to Ohio: Ohio and every swing state with a better chance of victory for Romney would be a narrow Electoral College win.
However, the distinction between base turnout and swing voter states illustrates the dilemma that Romney faced in his VP pick. A "safe" pick would not do much to swing the chances of victory in any state, since "safe" means "do no harm". Given the shaky relationship between Romney and the electoral base of the GOP, pursuing a more moderate Republican pick in order to "shake the etch-a-sketch" on the primary campaign and chase swing voters runs the risk of so alienating the base that is needed to win base turnout states and possibly treading water in the balanced states.
But ... then picking a radical reactionary ideologue has the flipside risk, of alienating persuadable swing voters ~ which might be right wing voters who are dubious on the extreme right, or center-right through to center-left voters who have bought the GOP spin that the Democratic policy stance is somewhere other than right wing to center-right, and so who imagine they are picking between center-right and center-left parties rather than picking between an extreme right wing and a center-right party.
And then ... Romney picked Ryan. Which makes it looked like turning his back on persuading persuadable voters and doubling down on a base-turnout strategy.
The Hidden Political Terrain of 2012
So, assume that the incumbent is sufficiently well financed and sufficiently competent to win the balance of persuadable voters against the ticket that includes the guy who wants to replace Medicare with Coupons for Grandma to buy private health insurance on a completely unregulated private health insurance market. That means that the Ryan pick is likely to put the "swing voter" swing states off the table.
And of course, picking Ryan also doubles down on the anti-immigrant Know-Nothing stance that Romney adopted as his only actual social policy position during the Republican primaries, so in addition to taking New Mexico off the table as a result of his primary run, the VP pick indicates that Nevada is being effectively given up on.
And what does Ryan conceding the "swing voter" and "Latino heavy" swing states off the table look like? It looks like this:
WI: 80.1% +10 ~ swing voter
NH: 73.5% +4 ~ swing voter
NV: 78.2% +6 ~ balanced / Latino Problem
IA: 63.6% +6 ~ swing voter
CO: 59.2% +9 ~ swing voter / Latino Problem
... 226/270 all swing voter / Latino Problem: +35
236/270 MN: 91.2% +10 ~ balanced
252/270 MI: 89.0% +16 ~ balanced
272/270 PA: 86.2% +20 ~ base turnout
275/270 OH: 67.9% +18 ~ balanced
288/270 VA: 66.2% +13 ~ base turnout
332/270 FL: 53.9% +29 ~ balanced
347/270 NC: 32.2% +15 ~ base turnout
Now, does that make sense? Well, there is one scenario in which it does, indeed, make sense, Nate Silver's model, like any data-intensive model, has a status-quo bias. That is, there is a strong inclination to treat the impact of things other than the things measured for the model ~ poll results, economic data, etc. ~ as being on average the same as before. There is some variability included, but the variance is around a mean value of "otherwise much as before".
Above is another map, from The Center for American Progress, on voter suppression efforts at the state level. We are currently facing the biggest wave of voter suppression effort since the establishment of Jim Crow laws in the former Confederate states. And while African-American voters are not a sole target, they are a substantial target.
The effort is not aimed at preventing all blacks from voting, but aimed at reducing the share of black voters in the general electorate. The same for elderly voters, college student voters, and poor voters. Those are the demographics least likely to have a driver's license, and largely consist of demographics most likely to vote Democratic. The likelihood is weakest in terms of elderly voters ... except when Social Security and Medicare are under direct threat, at which point many "Reagan Democrats" over 65 come home.
So just as the earlier version of Jim Crow included (rigged) literacy tests, the current wave of neo-Jim Crow laws focus on photo ID, with a state driver's license always included, and a passport is generally included as well.
Now, while Nate projects an 86% chance of Obama winning Pennsylvania, the projected margin of victory is 6 points: 53% to 47%. So Romney has to cover a roughly 6pt gap in Pennsylvania to squeak out a in Obama-leaning PA. In Obama-leaning Ohio, the projected margin is 2.4pts, 50.4% to 48%. In Obama-leaning Virginia, the projected margin is 1.8pts, 50.4% to 48.6%. In Florida, which Nate Silver has as a toss-up, the projected margin is 0.6pts, 49.9 % to 49.3%. So with just 2% net Democratic vote suppression, Ohio and Virginia are toss-ups, Florida is leaning to Romney.
And Pennsylvania? Is it possible to use voter suppression to cover a 6% gap? With the new Pennsylvania voter ID law just upheld in its court challenge, its possible that it might. As David Savage reports in classic he-said, she-said style from the Washington Bureau of the LA Times:
The two sides differed greatly on the potential effects of the law. In March, Republican leaders estimated about 1% of Pennsylvania's voters, or about 90,000 people, lacked the required ID cards. But in July, the state reported that about 9%, or more than 758,000 people, did not have a valid ID issued by the transportation agency. In heavily Democratic Philadelphia, 18% of registered voters did not have a current state driver's license, according to the state's data.If 90,000 voters are 1% of the electorate, then covering a 6% gap requires some combination of a difference in base enthusiasm and Democrat voter base suppression equalling 540,000+ votes. With 758,000 people without valid ID under the PA law, a 2:1 Democratic:Republican bias would cover 250,000 of those.
So, where does picking Ryan come into this? The base-turnout strategy requires you to get your base enthusiastic, especially in a state where your base is slightly smaller than the other establishment party's base, as in Pennsylvania.
Voter suppression? It requires much the same. You need the poll-workers from your party to be leaning as hard as they can to support the vote-suppression campaign on the ground. They need to be sticklers for every aspect of the voter ID law. And they need to be sticklers at the same time that confusing and misleading information about what ID is required is being spread in the media in the state. If its not required to offer a provisional ballot to someone without correct ID, they have to lean against their Democratic opposite number doing so: indeed, it is best if they badger their Democratic opposite number if their opposite number should be inclined to take time to clearly explain the provisional voting process, since two provisional votes cast without a follow-up visit to validate the vote is the same as one vote flipped to their side.
Effective voter suppression of the type required needs your base on the ground to be sufficiently enthusiastic about your campaign that they are completely plugged into getting the instructions and advice on effective voter suppression that will be available over back-channels, even as the media is being filled with vague, confusing and conflicting explanations to voters as to what is required.
And that's what the Ryan pick looks like, to me. A bid to win the support of the radical reactionary activist base of the Republican Party, by putting one of their own on the ticket.
It certainly cannot be an effort to persuade swinging voters, since the major legislative defeats for which Ryan has become a "player" inside the beltway involve proposing a big pile of massively unpopular policies, pushing them through the House on the back of the radical reactionary Republicans and those afraid of facing radical reactionary Republicans in primaries ~ also known jointly as the Republican Congressmen ~ to see them die in the Senate. That is not the man you pick to appeal to center-right independents unaffiliated to either party.
But as far as a base-turnout strategy, supercharged with voter suppression of the competing establishment party's base ... yeah, then the Ryan pick makes sense.
So my best guess is that is the way that the Romney campaign hopes to win the White House.
The Boss Says: John Henry