Monday, January 23, 2012

"There but For the Grace of God Go I"


On January 21, while the eyes of the political world were focused on the South Carolina primary, the Iowa Republican Party officially announced that Rick Santorum – notMitt Romney -- had won the Iowa caucuses.   Of course, for days we’d known that the vote total in Iowa was problematic, but the official results from election night (January 3), when Romney was declared the winner by 8 votes, remained in force until Saturday, January 21.  . The announcement of Romney’s victory was based, however, on the uncertified results from January 3.  This was where the problems began.   As both the2000 Presidential race in Florida and the 2008 Senate race between Al Franken and Norm Coleman in Minnesota should have taught us, the counting of votes does not end on election night.  It can take weeks for all the votes to be officially counted and certified.  During this period, vote totals change as elections officials discover uncounted ballots and add them to the vote totals.  In Iowa, it turned out that eight precincts never had their votes counted.  Vote totals were consequently revised; the resulting changes gave Santorum a 34-vote victory.

Most of the press commentary on Iowa’s vote counting gaffe has focused on the workings of the caucus system, the professionalism of the Iowa GOP, and questions whether Iowa’s standing as the first-in-the-nation state on the presidential calendar is threatened.  Some commentators have questioned the relevance of the Iowa caucuses, which don’t actually choose any delegates; they contend that the caucuses are largely a media invention with little practical impact except to drive out fringe presidential candidates.

While all of these are valid points, they all miss the big picture.   

What happened in Iowa was almost inevitable, given the closeness of the vote.  As I’ve argued in previous blog entries, the American electoral system is broken – or, at best, worn out and highly inefficient.  Vote-counting mistakes happen all the time.  Every election leaves a significant percentage of ballots uncounted, improperly counted more than once, or ignored for technical reasons.  The announced vote totals on election night are often little more than estimates with a margin of error that can be as large as 3 percent of the total vote.   For most elections, this imprecision does not matter.  Given that most candidates win by margins greater than the margin of error, the inaccuracies of the vote totals are usually irrelevant.  When, however, votes are close and every vote really does matter, then the system breaks down – and breaks down for all to see.

In 2001, Cathy Cox, then the chief election official of Georgia, remarked, “As the presidential election drama unfolded in Florida last November, one thought was foremost in my mind: there but for the grace of God go I. Because the truth is, if the presidential margin had been razor-thin in Georgia and if our election systems had undergone the same microscopic scrutiny that Florida endured, we would have fared no better. In many respects, we might have fared even worse.”  This was likely the thought of every election supervisor at the time.  Yet in most states, little was done to fix these problems.   Granted, we no longer vote on the old IBM punch-card machines of 2000 (which introduced us to the concepts of ‘hanging’ and’ pregnant’ chads).  Similarly, after a brief flirtation with electronic voting machines (which did not generate a paper record of the votes cast), most of us now vote on machines that provide a paper trail.  

Yet these changes address only mechanical problems.  They ignore the human element in vote-counting – including the potential for what I’ve termed administrative gerrymandering – manipulating the administrative procedures governing voting for partisan gain..  Yet even lacking election officials seeking to practice administrative gerrymandering, the failure to address the human element in running an election is a potentially catastrophic mistake.   The 2008 Senate race in Minnesota dramatized the ongoing inaccuracies and vulnerability of our electoral processes when confronting close votes.   Here the problem wasn’t chads or even paper trails, but the counting of absentee ballots – in particular the determination of which absentee ballots were valid and which were invalid, all for obscure technical reasons.  That this confusion arose in Minnesota showed the systemic nature of the vote-counting problem.  In 2008, Minnesota had some of the best-organized and best-- thought-out electoral procedures in the nation.  They had a non-partisan elections board to settle disputes and a state supreme court that also was nonpartisan in its makeup.  Yet, when facing a close vote, Minnesota turned out to be more like Florida than it wanted to be.  In fact, the only reason that the Minnesota mess was settled in a clear and generally-accepted fashion was that the Minnesota Supreme Court had the time -- over a month -- to decide the case, write their opinion, and get it right – time that no one in the Florida mess had.    Nonetheless, even following Minnesota’s 2008 electoral breakdown, little was done to fix the problem.

Instead, over the last few years, most efforts labeled as electoral reform have focused on preventing voter fraud through voter ID laws, reduced early-voting hours, and restrictive rules on third-party registration drives.  Leaving aside for the moment that, as the Brennan Center at New York University Law School notes, “Fraud by individual voters is both irrational and extremely rare” – which argues that most of these rule changes are politically driven for partisan advantage – the big problem with such ‘reform’ efforts is that they don’t address the real problems of vote-counting.   We are still as vulnerable today as we were in 2000 to close elections that demand extreme accuracy in vote-counting, an accuracy that at present we cannot ensure.  If anything, today’s problem is actually worse than it was in 2000, as candidates now come prepared on election eve with teams of lawyers ready to do battle in the courts for every vote.   

Add in administrative gerrymandering and the picture gets even more troubling.  Recall the notorious 1948 primary election for the Democratic nomination for a Texas U.S. Senate seat that pitted Lyndon B. Johnson against Coke Stevenson.  Johnson won by 202 votes from Jim Wells County – votes that came in six days late and were listed in alphabetical order [on this event, see the forthcoming article in The Review of Litigation by Josiah Daniel, “LBJ v. Coke Stevenson:  Lawyering for Control of the Disputed Texas Democratic Party Senatorial Primary Election of 1948”].  Another example is the case of the disputed ballots from the 1876 presidential election that ruined reputations, disclosed rampant corruption among Democrats and Republicans alike, and helped to bring a premature end to Reconstruction.  A third is the 2004 presidential election in Oho, where the state’s Republican secretary of state Ken Blackwell’s administrative rulings forced thousands of voters to fillout provisional ballots that never were counted.  I could list examples, both recent and historical, indefinitely.

Sadly, the real lesson of Iowa’s vote counting gaffe is that sooner or later a close vote in a key race is going to happen, and when it does, it will blow up in our face like an exploding cigar – leading election supervisors everywhere to think, “there but for the grace of God go I.”