Post Sandy Questions

Tropical Storm Sandy has created a disaster in New York and along the eastern seaboard.  The states that were affected all have to have their elections up and running on Tuesday; except for New Jersey, the states most directly affected are not states with early and absentee voting.

In New York, numerous subway tunnels were flooded and each tunnel has to be checked for debris by hand.  Power outages in the region are severe and so is flooding– CBS news has a state by state breakdown.  This disaster creates a two-part problem:

1.  The week before an election is critical for logistics.  All voting machines and tabulators have to be L&A tested.  Everything has to be sealed.  Supplies have to be packed.  Ballots have to be staged.  Polling place locations checked one last time.  Last minute training classes held.  All of this will now have to be packed into a shorter window and the potential for a major problem increases

2.  The storm may also create a series of political problems.  If the storm and its aftermath lowers turnout, it could not only affect the outcome of  close races but it could also affect the Presidential race rather directly.  Given how close national polling is between the two candidates, and that the storm hit Obama voters more directly, lower turnout among Democrats may not mean that Obama does not win the affected states but it could mean that we have a re-run of 2000 — an electoral college winner who does not win the popular vote.

FL Early Voting Update for 10/30

Here is the latest on Florida early voting, as of 9:30am on Tuesday, Oct. 30.  The statistics reflect close of voting yesterday.  For some reason, three counties have yet to provide an updated file:  Duval, Glades, and Union.  Duval is a significant omission, since the county’s web site reports over 20k early voters yesterday.  (There is no partisan breakdown of the vote, however.)

It continues to look like efforts to reduce the number of people voting early in-person has been a success.  With five more days of early voting left, just over 801k have voted by this mode.  In 2008, the cumulative number was 1.4m.  To match the 2008 numbers, the number of early voters will have to average 386k per day.  The record one-day early voting turnout is 295,405, set on the first day of early voting this year.

The updated spreadsheet can be found here:  https://www.dropbox.com/s/4nz7v0u1bitp3cx/fl_early_voting_comparison_20121030.xlsx

NYC Elections….How?

It is unclear to me how election officials will prepare for the election on November 6. With no power and parts of the city flooded, how exactly will they get polling places open, all the polling stations manned, all of the voting machines, voting equipment deployed? This will be the challenge of the election.

Another Election, Another Potential Disaster

As Hurricane Sandy approaches the coast of New Jersey and is causing havoc on the coast from North Carolina north, it is interesting to ask a simple question — what if this storm was hitting on November 5, not at the end of October?

The impact of such an event could be devastating to the ability of states to hold elections and to national politics.   Steven Huefner at the Moritz College of law wrote a nice legal analysis of some of the implications of Sandy on elections.   I want to make some of what he implied more specific and also raise some additional questions for consideration.

First, lets briefly consider some of the things likely to be wrought by Sandy.

Massive and lasting power outages.

  • No power means that DREs will only operate as long as their batteries last.  It also means that voters voting on paper ballots will not have the use of scanners to identify errors on their ballots.  But wait, there is more!
  • No power means no alarm clocks to wake up poll workers, no phones to call the custodian to open the school, and no lights in the school for voting.
  • No power means no call centers for when problems arise, no printing last minute changes to the voter registration roster, and limited use of those automatic poll books.

Evacuations and Traffic.

  • Some voters will literally not be able to vote because they will have been evacuated from their local polling place and there is no provision for remote voting.  Imagine if Long Island was under an evacuation notice; how would those voters vote?
  • Today in New York and DC, there is no transit.  How do people get to the polls where there is flooding, no power for traffic lights, and no public transport available?

The Horizontal State Problem and the Early Voting Problem

  • The horizontal state problem are best epitomized by Pennsylvania and New York.  Neither state has early voting and both have very strict absentee voting laws.  Hurricane Sandy hits tonight and there is no power on Election Day in Philadelphia or New York City.  People are warned to stay indoors because of downed power lines and flooding.  However, in the rest of both states, people can vote.  Such an event could systematically disenfranchise major metropolitan areas critical to determining who wins these states in Presidential and Senate races.  Also, who wins the contested House races in these localities.
  • The early voting problem is an extension of the horizontal state problem.  Imagine that North Carolina is severely hit — a state with extensive early voting and liberal absentee voting.  Does the system make any allowance for one voter having an easier time voting than the voter who wants to vote on Election Day?

Not A New Problem

Ever since 9/11, we have all been well-aware that disasters can completely disrupt an election.  However, Congress and state legislatures have avoided considering these contingencies.  Perhaps we should before we have a real constitutional crisis.

New report from VTP: Voting: What Has Changed, What Hasn’t, & What Needs Improvement

Recently the VTP released a new report, Voting: What Has Changed, What Hasn’t, & What Needs Improvement. It covers a lot of ground, examining what has happened in the administration of elections in the US since the 2000 presidential election — and providing a series of recommendations for issues that need research and improvement.

Update on Florida Early Voting

I’ve downloaded the most recent early voting updates.  Most counties are uploading their data in the wee hours of the morning following the previous day’s voting.  Two tiny counties (Glades and Lafayette)  have uploaded only one file, while the Polk County file was uploaded mid-evening last night.  I mention this only to note that there will likely be some inconsistencies in a few counties.  I am doing my best to make sure as much data as possible is being included in these updates.

Second day early voting dipped a bit (from 295k to 232k), for a total now of over 527k.  By the second day of voting in 2008, there had been 319k early ballots cast.  However, in 2008, with 7 days of early voting to go, there had been 1.2m ballots cast.  It remains to be seen whether Florida’s early voting totals in 2012 will equal its 2008 totals.

Democrats remain ahead in the early voting, but again, the margin is lower than in 2008.  Right now, Dems maintain a 49.6%-33.1% lead over the Reps.  In 2008, it was a 54.1%-30.9% lead on day two, or 53.8%-29.7% lead with 7 days of early voting to go.

The link to the new spreadsheet is here:  https://www.dropbox.com/s/ro6vfyu2pzgngra/fl_early_voting_comparison_20121029.xlsx

Florida early voting statistics

Here are the Florida early voting statistics, with comparisons to 2008.

It has to be said up front that it’s not clear what the proper comparison with 2008 is.  Changes to Florida’s early voting laws shortened the early voting period, from two weeks to eight days.  So, do we compare statistics from October 27, 2012 with the first day of early voting in 2008 (October 20, 2008) or with the eighth day from the end of the early voting period (October 26, 2008)?  My choice is the latter.  However, in writing about the start of early voting in Florida, it would also be useful to make comparisons with the first day of early voting in 2008.

In 2008, almost 158,000 people voted on the first day of early voting.  In 2012, that number was 295,000.  However, in 2008, by the time Florida had gotten to within eight days of the end of early voting, over 1 million people had voted.  So, you can see why it’s important to be clear about how the accounting is framed — are we comparing from the beginning of period, or the end of the period?

In 2008, the first day partisan breakdown was 56.6% Democrat, 29.0% Republican, 10.8% independent, and 3.6% other parties.  In 2012, the first day breakdown was 48.8% Democrat, 35.3% Republican, 13.1% independent, and 2.7% other parties.  In 2008, on the eighth day before the end of early voting, the partisan breakdown was 62.8% Democrat, 20.1% Republican, 13.8% independent, and 3.3% other.

Whichever way we look at it, the Democratic advantage in Florida early voting has diminished, from either a 27.5  or 42.7 percentage point advantage (depending on your perspective) down to a 13.5 percentage point advantage.

Let’s see what souls to the polls does.

I have posted the spreadsheet with the running comparisons at the following link:  https://www.dropbox.com/s/oqjmvykd7oa8j66/fl_early_voting_comparison_20121028.xlsx.

Election Fraud research from Fredrik Sjoberg

Fredrik Sjoberg recently sent along links to two quite interesting working papers from his research on election fraud.

Here are the papers, and their abstracts.

First, “Making Voters Count: Evidence from Field Experiments about the Efficacy of Domestic Election Observation.”

Here’s the abstract:

Elections are important because they hold the promise of empowering voters to hold leaders accountable. The sad reality, however, is that voters in less than democratic states often are marginalized because of widespread election fraud. Field experiments in three different countries are here used to show that high-quality civil society observers can reduce fraud on election day. The results also confirm that all regimes are not equally sensitive to such interventions. For the first time new fraud forensics techniques are used to examine observer effects. I argue that a reduction in detectable fraud forces authorities to engage with the electorate more directly, instead of focusing their efforts on bureaucratically manipulating the outcome. It is suggested that when faced with monitoring, autocrats substitute election fraud with other forms of manipulation, in the form of vote buying and intimidation. This in itself constitutes a perverse form of empowerment of voters, perverse since the process continues to be both un-free and unfair.

Second, “Autocratic Adaptation: The Strategic Use of Transparency and The Persistence of Election Fraud.”

And the second paper’s abstract:

Why would an autocrat want, or at least make it appear to want, to reduce election fraud? In recent years, non-democratic rulers have surprisingly begun to embrace fraud-reducing technologies, like web cameras or transparent ballot boxes. The reason for this is not found in international norms or domestic conditions for post-electoral protest, but rather in the null effect on the ruling party vote share. With the help of new fraud identification techniques, I argue that the installation of web cameras in polling stations changes how fraud is conducted. Web cameras do not reduce fraud, but rather make certain blatant forms of fraud, like ballot box stuffing, more costly. Autocrats then substitute for other types of fraud, such as fabricating vote count outside the view of the cameras, in order to secure electoral victory. Overall, this paper identifies this compensation mechanism where incumbents are able to prevent vote share losses, while contributing a veneer of legitimacy by self-initiating anti-fraud measures.

Both of these are interesting papers to those who study election fraud and electoral monitoring.

Virtual Issue of Political Analysis: Election Fraud and Electoral Integrity

Political Analysis has just published a virtual issue on Election Fraud and Electoral Integrity, edited by Ines Levin and myself.

The virtual issue contains a number of papers published recently in Political Analysis, on the forensics of election fraud and on how to study electoral outcomes.

The virtual issue’s papers are freely available for a limited time, as is the the introduction that Ines and I wrote.

Interesting talk by David Chaum in Cambridge next week

For anyone in the Boston area, David Chaum is giving what promises to be a very interesting talk next week at Harvard on “Random-Sample Elections.” Anyone who knows David or has seen him talk knows that he’s an interesting, engaging guy. This is an event not to be missed.

When: Thursday, Oct. 4, 4:00-5:30 pm
Where: Harvard, Dworkin G125
What: The Computer Science Colloquium Series

This link has more information.