Skip to main content

U.S. Republican presidential candidate and former Massachusetts governor Mitt Romney speaks to supporters during a campaign rally in Pascagoula, Miss., on March 8, 2012.DAN ANDERSON/Reuters

Republicans are spinning that their long and ugly nomination fight will have no impact on their chances in the 2012 presidential election.

"Clinton and Obama went after each other until June in 2008, and it certainly didn't affect the president's chances going forward when he won that November," Virginia Governor Bob McDonnell says.

Really? The math says differently. Here's how.

There have been 20 total nominations since primaries became the preferred method to select nominees in 1972, 10 Democrat and 10 Republican.

Assign each primary a score between one and six for how long and ugly their primary season was, with one being literally bloody and six being uncontested. Then create a single variable for how much worse one party's nomination process was than the other party's by simply subtracting the Republican score from the Democrat score in any given year. Finally, use regression analysis to see if there is a relationship between our score for how much better or worse the Republican primary was than the Democrat, and how the Republicans did in November.

---

STEP 1: SCORING THE PAST NOMINATIONS

Let's start with coding the individual processes for each of the 20 nominations.

(We are starting in 1972 because that is the first U.S. presidential election cycle to take place after primaries became the system for choosing party nominees. Prior to 1972, delegates to the national conventions were selected by regional party bosses in most jurisdictions. As a result, conventions in that era were more often brokered between factions on the floor. After 1972, nominees began to routinely sew up the party nomination before the final primary was even held.)

We will assign each of the 20 nominations a score between one and six.

1 – If the nomination is undecided all the way to a vote on the floor.

2 – If the presumptive nominee is openly challenged on the floor.

3 – If the nominee secures presumptive victory between Super Tuesday and the end of the primaries.

4 – If the nominee secures presumptive victory at or before Super Tuesday.

5 – If the clear front-runner faces real challengers which are easily set aside.

6 – If the clear front-runner faces no real challengers.

Now let's go through the primaries an assign each one a code.

1972

Democrat:This was "perhaps the most contentious convention in the history of the Democratic Party since 1924." The traditional Democrat coalition of labour unions, city bosses and Southerners narrowly lost control of the Convention on delegate seating rules, allowing Senator George McGovern to get the nomination in a chaotic all-night disaster. One is the lowest score, and it deserves it.

Republican: Richard Nixon faced a couple moderately serious challengers but won the support of all but one delegate at the first scripted set-piece convention. I'm giving this a five.

Result: Nixon beat McGovern in one of the largest landslides on record with 60.7 per cent of the vote.

1976

Democrat: Governor Jimmy Carter beat Senator Scoop Jackson in Pennsylvania and essentially wrapped up the nomination. There were late challenges by Western liberals but the convention was a love-in for Carter. I gave this a three.

Republican: Governor Ronald Reagan challenged President Ford at the convention. However, Ford had a clear majority before the first floor vote after Reagan fumbled the days before the convention with the announcement of a liberal Republican for his VP candidate. The choice lost him conservative support while gaining him nothing with moderates and liberals. This is a two.

Result: Carter narrowly beat Ford.

1980

Democrat: Senator Ted Kennedy viciously attacked President Carter all the way to one of the nastier conventions on record, refusing to drop out until the day before nominations. This is a two.

Republican: Reagan was the presumptive nominee until he unexpectedly lost to George H.W. Bush in Iowa. Victories in New Hampshire and South Carolina put Reagan back on top. This is a four.

Result: Reagan won decisively in the electoral college but only got 50.7 per cent of the popular vote.

1984

Democrat: Senator Gary Hart challenged former vice-president Walter Mondale all the way to the convention, where Mondale had a clear majority thanks to the support of the unpledged "Super-delegates" of the party establishment. This is a two.

Republican: Reagan was unchallenged as sitting President. This is a six.

Result: Reagan won every state except Washington, DC, and Mondale's home of Minnesota.

1988

Democrat: Governor Michael Dukakis wrapped it up post-Super Tuesday after Rev. Jesse Jackson and Senator Al Gore failed to build momentum. This is a three.

Republican: George H.W. Bush was endorsed by Reagan and overcame a loss in New Hampshire to wrap it up early. This is a five.

Result: Bush won convincingly with 53.4 per cent of the vote.

1992

Democrat: Bill Clinton quickly built a majority coalition, sweeping Super Tuesday, but faced a stubborn refusal to concede from Jerry Brown. This is a three.

Republican: President Bush was challenged by Pat Buchanan. Bush never lost a state. This is a five.

Result: Bush was demolished, getting just 37.5 per cent of the vote, the lowest for a Republican in our series.

(Of all our findings, this one appears to be an outlier. It may be the case that this challenge by Buchanan was more damaging than the numbers would show. There may be an argument that challenging a president is a sign of greater party disunity, which would lead to lower scores for the Republicans in 1976, Democrats in 1980 and Republicans in 1992.)

1996

Democrat: Clinton was unopposed. This is a six.

Republican: Bob Dole was the presumed nominee, faced some challenges, but wrapped up the nomination early and clean. This is a four.

Result: Clinton decisively beat Dole.

2000

Democrat: New Jersey Senator Bill Bradley challenged vice-president Al Gore for the nomination, but Gore won every state. This is a five.

Republican: Texas Governor George W. Bush was the early front-runner and overcame a loss on John McCain in New Hampshire to wrap it up early. This is a 4.

Result: This was the closest election in U.S. history, with Bush edging Gore by a few hundred votes in Florida to win the electoral college while losing the popular vote.

2004

Democrat: John Kerry slowly emerged as the front-runner after Iowa and then quickly locked up the nomination by Super Tuesday. This is a four.

Republican: President Bush was unopposed. This is a six.

Result: Bush narrowly won re-election in the electoral college by winning Ohio, but got a majority of the popular vote with 50.7 per cent.

2008

Democrat: Barack Obama and Hillary Clinton contested primaries into June, but Obama clinched the nomination well before the convention. This is a three.

Republican: John McCain was the presumptive nominee after Super Tuesday, despite Mike Huckabee running a faint hope campaign for several weeks after the result was clear. This is a four.

Result: Obama beat McCain decisively.

---

STEP 2: COMPARING THE REPUBLICANS AND DEMOCRATS

Now that we have our scores for each primary, we can compare the nomination process of each party each year and create a value, which is the "bonus" that the Republicans got each presidential election for their nominating process. If the number is positive, the Republicans wrapped up their nomination with more ease and cleanliness than the Democrats. If the value is negative, the Democrats had an easier time selecting a nominee.

Year – Democrat / Republican / Bonus

1972 – Dem. 1 / GOP 5 / Bonus 4

1976 – Dem. 3 / GOP 2 / Bonus -1

1980 – Dem. 2 / GOP 4 / Bonus 2

1984 – Dem. 2 / GOP 6 / Bonus 4

1988 – Dem. 3 / GOP 5 / Bonus 2

1992 – Dem. 3 / GOP 5 / Bonus 2

1996 – Dem. 6 / GOP 4 / Bonus -2

2000 – Dem. 5 / GOP 4 / Bonus -1

2004 – Dem. 4 / GOP 6 / Bonus 2

2008 – Dem. 3 / GOP 4 / Bonus 1

Just looking at the number, some patterns emerge. We see big advantages for the Republicans in 1972 and 1984, years that were Republican landslides. We see slight Democrat advantages in 1976 and 2000, very close years. We see a stronger Democrat advantage when that party had a decisive win in 1996.

STEP 3: REGRESSION

Regression is a mathematical way to test the relationship between two variables.

Basically, we set up a grid. We plot each of our election years according to two variables. Along the bottom x-axis, we find the score for the net Republican bonus for the primary process, the score we just assigned each election. Along the side, on the y-axis, we find either the popular vote for the Republicans in the general election for our first regression, or the Republican electoral college vote in the second regression.

Excel will then take all these points and plot a straight line that finds the least distance between itself and the points of data. Depending on how well the line fits the data, the explanatory power of the model is found in the R-squared and the significance.

If we run this regression, we find a number of things.

First, there is a significant relationship between the nominating bonus and the popular vote (significance 0.036) and the electoral college (significance 0.016). So we can discard the suggestion that any relationship between the two is just random chance or background noise.

Second, the R squared with the popular vote is 0.43 so it is explaining about four-tenths of the behaviour of the electorate when they vote.

Third, the relationship is even stronger when we regress the relative pain of each party's nominating process against the electoral college results. There we see an R squared of 0.53, so about half of the variation in the electoral college result is related to the variation in the nominating process.

---

STEP 4: ANALYSIS

So what does it all mean?

For starters, this model doesn't explain nearly everything in the general election. There are years where the party with the qualitatively "better" primary process loses the general election. 1992, 2000 and 2008 were like this.

However, there are broader issues at play in the general election dynamic: state of the economy, war, voter exhaustion with one party in the Presidency. The partisan nominating process is one factor that may be turning a landslide into a tight contest (2000, 1992) or a tight contest into a landslide (1972, 1984).

Rather than look just at outcome, it is better to look for relationships between the nominating process and the result in the general election.

At the very least, we can say this: There appears to be a significant impact on voting behaviour and electoral college results in U.S. presidential elections when one party has a smoother nominating process than the other.

If this will turn a close election into a landslide or make the difference between a Democrat and Republican victory this November depends on how ugly the Republican contest gets and the myriad of broader factors from the economy to Iran to debate performance that are yet to be decided.

So far, Mitt Romney still looks likely to sew up the nomination before the convention, earning a 3 compared to Obama's 6.

But bad blood with Santorum, Gingrich or Ron Paul could turn the convention into a mess, particularly if Romney can't assemble a clear majority of delegates before the last primary.

---

STEP 5: DISCUSSION

There are some important caveats.

This model is based on qualitative assessments of each primary season, which could influence in the outcome. Was the 1984 Democrat process really as destructive as the 1976 Republican showdown? Moving a few of the variables around can influence the outcome enough to make it less significant. The weakness makes it dangerous to draw out "rules" or other general conclusions, but it does show a relationship between long, drawn out primaries and poor electoral results for that party.

Much of this is a question of causation. I don't think the length of the primary process is actually directly influencing the vote, as if the date on the calendar of some nominee locking up a majority of delegates holds much importance with voters. Instead, the length of the primary process is itself reflecting the challenges within each party, and it is those challenges that are then impacting the general election result.

Some of the lowest scores are associated with parties going through difficult ideological breaks. The 1972 Democrats were split between new Left activists and comparatively conservative union and southern factions. The 1976 Republicans were openly split between Reagan's conservatives and Ford's moderates. The 1984 Democrat convention was a fight between the old liberal union establishment from the Roosevelt era and the "new Democrats" who would take power with Clinton. The 1992 Republicans did exhibit signs of a split between hard-core conservatives revolting under Buchanan and the party establishment of more centre-right conservatives who rallied to defend President Bush.

However, some of these are simply personality fights, particularly the 2008 Democrats. There was little ideologically different about Barack Obama and Hillary Clinton, beyond how they would have voted on the Iraq War. The issue there was more about the complex coalition politics of the Democratic Party.

Another example of this is the 1980 Democratic primary, which was to some extent an exercise in hubris by Senator Kennedy in trying to seize the presidential nomination he spurned after Chappaquiddick made it harder for him to win in 1972 and 1976.

Some are issues of the quality of the nominees themselves. The 2004 Democrat field was particularly weak, with Kerry on emerging as the front-runner after winning both Iowa and New Hampshire and even then, failing to positively energize a Democratic Party more desperate to beat Bush than govern. Similarly, the 1996 Republican field was surprisingly weak given Clinton's standing in the polls in 2004 and early 2005. This may be a function of presidential nominees selecting unknown and somewhat unqualified candidates in an attempt to goose their poll numbers, resulting in fewer "presidential nominee in waiting" level contenders four years later. Dan Quayle, John Edward and Sarah Palin are examples of this recent phenomenon.

There could be an issue of causality here, where the splits within parties are separately impacting their vote in the general election, the quality of the nominees, and the ability to wrap up nominations quietly.

What we could be seeing is that parties that don't know who they are have a hard time nominating and a hard time winning.

Whether the long nomination process impacts the election, or the splits in the party impact both the nomination and the election, it is clear than the Republican will be carrying something of a primary-imposed anchor going into the general election this November.

Interact with The Globe