Home > News > Earning More Seats with Fewer Votes: Why the 1996 House Election Results are Not Necessarily a Good Analogy for 2012
194 views 10 min 0 Comment

Earning More Seats with Fewer Votes: Why the 1996 House Election Results are Not Necessarily a Good Analogy for 2012

- December 6, 2012

The following is a guest post from Yale University political scientist David Mayhew, the author of Congress: The Electoral Connection.


In last month’s election, the Democrats won more popular votes for the House but the Republicans kept the House. Has this kind of disparity happened before? People have been searching backwards, and the 1996 election has emerged as one candidate. Is the case for 1996 good? It is a complicated story. The answer seems to be officially yes, but really no.

The official source for the vote is Statistics of the Presidential and Congressional Election of November 5, 1996, compiled by Robin H. Carle, Clerk of the House of Representatives.

“Recapitulation of Votes Cast,” pp. 81-82, reports 43,121,000 votes nationwide in the Republican column, 43,394,000 in the Democratic column (to the nearest thousand).  That is a Democratic edge of 273,000, or 50.16% of the major-party vote.

For counting votes, the House Clerk used decision rules that made sense and were no doubt time-tested.   One of them was, for any district, to count toward the national Democratic or Republican totals the popular votes cast for any winner whose vote count was officially available, yet to score a zero for any party, including as relevant the Democratic or Republican Party, that didn’t field an opposing candidate.  There were 13 such elections that lacked major-party opponents in 1996, of which six were won by Republicans, seven by Democrats.  For example, two South Carolina districts won by Republicans were scored as 138,000 and 158,000 for the GOP winners, zero and zero for the Democrats.  Two West Virginia districts won by Democrats were scored as 171,000 and 146,000 for the Democratic winners, zero and zero for the Republicans.  These decisions were standard.  A party that didn’t put up a candidate wasn’t credited with any votes.  A party’s winner who had votes officially cast and reported for him/her got credited with them notwithstanding a lack of major-party opposition.

So far so good.  But there were several anomalies in 1996.  These were cases that seem to deserve asterisks for one reason or another.  Almost all of them point to under-assessments of pro-GOP sentiment in the House elections of that year.  Taken as a whole, they pretty much undermine the image of a national Democratic edge, notwithstanding the official statistics.

Here are the anomalies:

  • Louisiana elected seven MCs in 1996, five Republicans and two Democrats.  But the state’s vote was recorded as 136,000 Republican, 339,000 Democratic.  How could this dissonance have happened?   On exhibit is the state’s nonpartisan jungle primary.  Here is a House Clerk’s note regarding Louisiana in 1996:  “A candidate who is unopposed is declared elected by the people and his/her name shall not appear on the ballot in either the primary or general election.”  Thus Republican winners Bob Livingston and Billy Tauzin were credited with no votes at all in the Clerk’s report.   Another Clerk’s note:  “A candidate who receives a majority of the votes in the primary election is declared elected.”  Instances like these also resulted in zeros.  Thus Democratic winner William Jefferson, and Republican winners Jim McCrery and Richard Baker, were credited with no votes at all in 1996.  Beyond these five instances, the election in the Louisiana 5th district was normal, but the 7th district brought another anomaly.  In today’s terms, it was a Berman-Sherman contest (there was only one more of these in the country in 1996).  Two Democrats defeated everybody else in the primary, thus advancing to the general election where they shared 242,000 votes, which became the Democratic score for that district in the House Clerk’s document.  (According to Michael Barone and Grant Ujifusa, The Almanac of American Politics 1998, the earlier primary in the 7th district, in which turnout was lower, had drawn several unsuccessful non-top-two candidates, including three GOP candidates who shared 59,000 votes.)
  • In the Florida 4th district, Republican winner Tillie Fowler ran uncontested and was credited by the House Clerk with zero votes.  A Clerk’s note:  “According to Florida law, the names of those with no opposition are not printed on the ballot.”

  • Connecticut is a case where the pro-Democratic sentiment of voters, so to speak, was underassessed in the Clerk’s report.  The state allows candidates to run on more than one party line on the November ballot.  In 1996, the state’s Democratic House candidates drew 49,000 votes on a third-party line.  The Republican candidates drew none.  In nonofficial reportage, such third-party votes are ordinarily counted toward major-party candidate totals.  Barone and Ujifusa do that, for example.  But the House Clerk did not do that in 1996.  These 49,000 votes do not appear in the Clerk’s national House totals for Democrats.
  • New York allows cross-filing, too.  In 1996, the state’s Democratic House candidates racked up an extra 122,000 votes on other party lines.   The state’s Republican candidates racked up an extra 311,000 votes (mainly on the Conservative line), a considerably larger showing.  These figures did not count toward the national Democratic and Republican totals in the Clerk’s report.
  • Texas was a mess in 1996.  Courts intruded into the House elections late, causing 13 of the state’s 30 districts to hold nonpartisan jungle primaries on the November election-day ballot, with ensuing December runoffs where needed.   In 10 0f these 13 cases, the House Clerk sensibly used the votes cast in the jungle primary as entries in the national Democratic and Republican totals.  In any of these districts, all votes cast for any of the set of candidates listing themselves on the primary ballot as Democrats went toward the national Democratic vote total; similarly for the Republicans.  These ten contests produced majority winners on November election day.  Yet the other three contests went to a December runoff where turnout was much lower.  In these cases, the House Clerk applied the December results toward the official national totals.  In the November jungle primary, the total vote across these three districts (not used by the Clerk) had been 205,000 Democratic, 307,000 Republican.  In the December runoffs, the total vote (used by the Clerk) was 88,000 Democratic, 126,000 Republican.  In short, the Republican vote fell by 181,000 in December; the Democratic vote fell by 117,000.  (One of the runoffs featured the second Berman-Sherman contest of 1996, between two Republicans; left behind in this instance were 38,000 votes cast for losing Democrats in that district’s November jungle primary.)
  • There was an oddity in the Missouri 8th district.  Republican incumbent Bill Emerson died in June 1996.  That was apparently too late for his wife Jo Ann Emerson to file for the Republican nomination.  She ran as an Independent in November.  She was obviously the favorite of Republican voters and a sure thing to organize with the GOP in Washington.  The vote in the November election was 112,000 for Jo Ann Emerson as Independent, 23,000 for the regularly nominated Republican candidate, 83,000 for the Democratic candidate.  Necessarily, the House Clerk scored this election as 23,000 for the Republicans, 83,000 for the Democrats.
  • Finally, there was Vermont’s at-large seat, which Independent Bernard Sanders won with 141,000 votes over a Democrat with 24,000 and a Republican with 83,000. The House Clerk counted 24,000 for the Democrats, 83,000 for the Republicans.   It is obvious that many Democratic voters supported Sanders for strategic or other reasons.  Yet, it was also clear that Sanders was a genuine Independent not headed for instant post-election conversion into a generic Democrat.

Here is the gist.  The Republicans probably “lost” something like a net 800,000 vote edge by virtue of the processes taken together in Louisiana, Florida, Connecticut, New York, and Texas.  (This estimate disregards Missouri and Vermont.)  The key ingredient is the zero votes for the five GOP winners in Louisiana and Florida, versus the zero votes for the one Democratic winner in Louisiana.  A House victory is ordinarily worth 150,000 votes or so.