How Hurricane Sandy Could Seriously Disrupt the Election
Hurricane Sandy's development was impeded a bit by Cuba -- it likely would be a Category 3+ right now, if a mountainous land mass had not intervened -- but it remains a powerful Category 2 hurricane, with 105 mph winds and a barometric pressure of 963 mb, as it moves into the Bahamas. And the National Hurricane Center's 11:00 AM forecast has shifted west, following the consensus of the computer models, to squarely target the U.S. East Coast:
You should not focus on the exact forecast track, particularly in the latter parts of the forecast, as the details are still very uncertain at this point. Everyone in the "cone of uncertainty," from North Carolina to New England, should begin preparing for the possibility of hurricane-like conditions late this weekend or early next week. Also -- another reason to focus on the cone, not the exact track -- Sandy will have a very large wind field, with bad weather quite far away from the center. So a large area is going to be impacted by this storm.
[NOTE: For my latest updates on Sandy, follow me on Twitter.]
Widespread, long-lasting power outages are almost a given with a storm like this. Wind damage will be significant, particularly with leaves still on the trees (as opposed to winter Nor'easters), and particularly with heavy wind over such a wide area, which means affected areas will take a prolonged battering. Severe coastal flooding is also a serious threat, primarily north and east of wherever the storm's center makes landfall. (Monday is astronomical high tide, making things worse.) Severe inland flooding from heavy rainfall is another big potential problem. Places where those two phenomena can happen simultaneously -- for instance, the mouth of the Delaware River, where it flows into Delaware Bay -- are especially at risk, if the center comes in just to their south. Which, as it happens, is exactly what the 00z run of the European model predicts for Delaware Bay:
Yikes. I don't want people to #PANIC about a specific model or a particular model run; any individual scenario is unlikely at this point, because there is still a lot of time and a lot of uncertainty. However, that particular scenario -- what Joe Bastardi calls the "Philadelphia Story" scenario -- is one of many that's "in play" at the moment, which is distressing. Likewise, a New York harbor nightmare is another scenario that's in play.
Aside from all the "usual" impacts of a hurricane transitioning into an uncommonly severe hybrid coastal storm and impacting the Megalopolis, this storm is going to hit one week before a presidential election, which raises a whole host of additional concerns. I discussed some of these yesterday, including the possibility that the occurrence of a national emergency could alter the dynamic of the campaign in its final weeks. But right now, I want to focus on the procedural issues, the impacts on the actual conduct of the election itself. For instance:
• Additional chaos and "irregularities" on Election Day due to lost "prep" time. I spoke this morning with my father, a retired elections bureaucrat in Connecticut, and he made the excellent point that the week before the election is very busy for folks like him in his old job, and for registrars of voters, town clerks and the like. They're testing voting machines, printing ballots or other critical papers, and doing all sorts of other mundane tasks that are critical to assuring a smooth Election Day. If the impact of the storm wipes out all or part of that critical "prep week," then even if things are relatively "back to normal" by Election Day (by no means a given; see below), there would likely be an invisible storm impact in the form of additional chaos, "irregularities" and all manner of disruptions at the polls -- failed voting machines, missing ballots, etc. -- simply because the officials had to cut short their preparation, so more mistakes will inevitably happen. This, in turn, will increase the already-high likelihood of cries of fraud (from Republicans) and suppression/disenfranchisement (from Democrats) in the event of any remotely close outcome. Basically, Sandy is likely to make an already highly charged atmosphere surrounding the conduct of the election even moreso.
• Lower turnout impacting various elections, and possibly giving Mitt Romney the national popular vote. If power remains out, some roads remain impassable, and people's lives are generally disrupted in a myriad of ways a week after the storm in heavily affected regions (all of which seems entirely plausible), it is likely that lower turnout would result. This is especially true in states where the presidential race is not hotly contested -- basically, the entire region except for New Hampshire, Pennsylvania, Virginia, and Ohio -- so voters may be less-motivated to vote, and the campaigns would certainly be less-motivated to pull out all the stops to help them do so amid difficult circumstances. If turnout is particularly affected in Democratic strongholds like New York, much of New England, Maryland, etc., it could depress President Obama's national popular vote numbers, and help make an already much-discussed scenario even more likely: Obama wins the electoral vote and thus the presidency, but Romney wins the national popular vote (essentially, a reverse of the Gore/Bush result in 2000).
Of course, lower turnout in Virginia, Pennsylvania, New Hampshire, and/or -- above all -- Ohio, would be far more significant, as it could impact the Electoral College outcome. The nature of such an impact is a little hard to gauge, though. Obviously, it depends on which regions within those states are hardest hit. For instance, if Philadelphia is ravaged by Sandy, it could be an enormous problem for Obama, turning a state that appears to favor him into a real tossup, and perhaps a national tipping-point state. Conversely, if western Pennsylvania is hit especially hard -- whether by copious rain and strong winds, or possibly by heavy snow (!) -- it could hurt Romney, and put the Keystone State far out of his reach. Similar analyses would hold in other states where the voting patterns are very regional.
More broadly, due to the "enthusiasm gap," among other factors, low turnout probably favors the Republicans, at least in states without early voting. In states with widespread early voting (like Ohio), low Election Day turnout probably favors whichever party is doing better in early voting. I have no idea what the impact would be on key Senate races in places like Connecticut, Massachusetts and Virginia (and to a lesser extent the marginally close races in Pennsylvania, Ohio and Maine); it presumably would depend in part on the local "enthusiasm gap" there vis a vis the Senate races, but I don't know what those numbers look like.
• Widespread power outages causing disruptions in voting. This one seems highly likely to occur, in some form or another, in various places. Think back to Hurricane Isaac, or the summer derecho, or last year's Halloween snowstorm. Widespread power failures don't just get fixed overnight. A week after the storm, significant portions of the hardest-hit areas will likely still be without power. What will this mean?
Well, the first question to ask is, what kind of voting machines do the potentially affected areas use? VerifiedVoting.org has a helpful map showing the types of systems in use, state-by-state -- and county-by-county, if you click on an individual state (for instance, Pennsylvania, Virginia and Ohio). You can even click on the county and find out the specific manufacturer of the machine they use there.
As it happens, all of New England and New York apparently uses "optical scan" paper ballots, which means the process of voting can theoretically take place without electrical power (although it may be difficult to see the ballot without lights!). Typically, as I understand it, voters submit their ballots by feeding them into the optical scan machine, which may or may not have enough battery power to continue working in such an environment (I don't know that answer, and suspect it varies from manufacturer to manufacturer). But, worst-case scenario, presumably the ballots can be placed in a box, then submitted to the optical scan machine later, once power is back on. This could severely delay the counting of ballots (more on that later), but would not prevent people from voting.
More potentially problematic are the "DRE" (direct-electronic recording) machines, usually called "touch-screens." Those are used throughout all of New Jersey, Maryland and Delaware, and most of Pennsylvania, Virginia, West Virginia and Ohio. For instance, Pennsylvania's Philadelphia and Delaware counties -- which would be very hard-hit in the Bastardi/EURO scenario -- both use Danaher Controls' Shouptronic 1242 model. Good news there: according to a report (PDF) by the Department of Elections for New Castle County, Delaware, the Shouptronic 1242 is supposed to keep functioning even when the lights are off:
In the event of a power outage the Electronic 1242 is equipped with a battery backup adequate to maintain normal operation for up to 16 hours. The internal power supply maintains the battery charge when power line voltage is normal to assure that the backup battery is fully charged and immediately ready for use if the AC power fails.
Assuming the battery is fully charged before the power goes out, and assuming the machine remains off until Election Day, that would seem to be good enough to keep things going. But I have no idea (and don't have to check) whether every DRE system is equally robust, or if those 16-hour assurances have been fully tested, etc. And in the event of widespread machine failure, don't just assume "oh well, people will just use paper ballots instead." It's not that simple. Election officials do not routinely print enough paper/provisional ballots for the entire electorate to use them. If the regular voting machines don't work, there will be mass chaos at the polls -- and probably a great deal of disenfranchisement -- unless these places have a very good contingency plan, and put it into place well in advance.
Moreover, my point about polling places lacking electricity and lighting is not a trivial concern, even if the machines themselves are workable. The many thousands of polling places in school gymnasiums, local libraries and whatnot aren't typically hardened against power outages, and likely lack a contingency plan for bringing in generators and such on short notice. It is easy to imagine last-minute changes in polling place locations, many voters trying to vote in the "wrong" precinct because theirs lacks power, etc. At a minimum, I would anticipate a whole lot of chaos, confusion, and a higher-than-usual number of provisional ballots being cast.
• Widespread power outages causing major delays in vote-counting. Casting ballots is one thing; counting them is another; and communicating those vote counts to a central location is yet another. We are used to rapid vote counts filling our screens on Election Night, telling us who the winner is (albeit based on unofficial returns) within a matter of hours, except when the result is razor-close like in 2000. However, such speedy counting depends on thousands of local polling places having the ability to tally and transmit their vote totals shortly after the polls close. Without diving too far into the weeds again, it's easy to imagine widespread power outages causing all sorts of disruptions to this process. Consequently, we could easily see a number of states with uncommonly slow vote counts. If the slow-counting state is, say, Vermont or Maryland, nobody will much care nationally; if it's, say, New Hampshire or Virginia, it could make for a frustrating night.