Sunday, July 15, 2012

Who Won the Week (July 8-14)


The hits this week just kept coming for Romney camp.   The debate over outsourcing has not gone away; in fact, it has intensified.  President Obama’s team has begun to make the case that Romney doesn’t care about jobs, just the profits.  Romney, in what could be called an “I am rubber, you are glue” moment, has labeled President Obama as the “Outsourcer in Chief”. 

In the end, the advantage on this argument goes to Obama, as the Romney campaign is only further adding legs to the outsourcing story.  They also did this two weeks ago when it was revealed that Bain invested in companies that pioneered outsourcing; the Romney camp, instead of refuting the claim, said there was a difference between “outsourcing” and “offshoring”.  Either way, they are debating semantics, and the debate only helps the President.

All of that would be small potatoes if there wasn’t another bit of damaging news this week for Mr. Romney.  The Boston Globe came out with a piece showing that Romney had, in fact, been actively involved in Bain Capital long after the candidate had originally said. 

The Obama campaign was quick to pounce on this information, saying that Romney had either committed a felony by lying on Federal disclosure forms or that Mr. Romney had lied to the American people. 

Which brings us to the tax problem.  In the primary campaign, Mitt Romney (after much pressure) had released his 2010 tax returns and had released a partial return for 2011.  His campaign had believed that in doing so, they would take the issue off the table for the general election. 

Boy, were they wrong.  A growing chorus of Democrats, the media, and even many Republicans are demanding that Mitt release more than what he has already.  On Friday, Mr. Romney answered those calls by refusing to release anything more than he already had, again, playing right into the President’s hands.

In effect, the Romney campaign has likely concluded that there is more potential damage in making public his taxes than suffering what will probably be the constant attacks on him for not releasing them. 

On an only slightly related note, the Romney campaign on Thursday floated Condoleezza Rice’s name as being on the candidate’s short list for Vice President.  The move, however, was largely seen as a way to distract from the bad news cycle that had been manifesting this week from the Bain and tax issues.

However the news wasn’t only bad for Mitt this week.  On Monday, it was revealed that President Obama had his best fundraising month yet… except that he still fell far behind Mr. Romney.  How far?  Romney raised $106 M to Obama’s $71 M.  If this pace keeps up through the election, it’s conceivable that Romney could actually be better funded than the President, something that looked to be impossible just a couple months ago. 

Polling:  The President’s lead is relatively unchanged since last week (+2.4 according to RCP).  However, as we are getting closer to November, the statewide polls will matter more.  In that regard, the President seemed to have a very good week in polling.  Two polls in Wisconsin gave the President a 6 and 8-point lead, while a PPP poll in Virginia put the race at +8 Obama.  Finally, a Rasmussen poll found Florida to be a tossup, which can be seen as good news for the President as Rasmussen polls tend to lean towards the GOP.

Who Won the Week?  President Obama

Friday, July 13, 2012

Does the Home-State Matter?

What made me initially ask myself this question was the thought of which state Mitt Romney actually will call "home" for the 2012 election.  Since his campaign is centered in Boston, and he was once Governor of the Bay State, it would seem obvious that Mr. Romney would choose Massachusetts to be the state he calls home and appears next to his name on the ballot.  

However, he certainly didn't have to choose Massachusetts; he had plenty of other options, and he perhaps would have been better suited to choose one of those.  For example, Mitt Romney grew up and was raised in Michigan, and he clearly is nostalgic for the height of the trees there.  Of course, if he were to choose Michigan as his home-state, he would have to acquire a house there, though I doubt that would be much of a problem for Mr. Romney.  

Besides Michigan and Massachusetts, Mitt Romney has a home in La Jolla, California, a ski lodge in Utah, and two lakeside homes in New Hampshire.  A couple of these options seem to represent zero political gain.  However, Michigan and New Hampshire are both currently considered swing states and could each potentially shift the outcome of the election.

But I am getting ahead of myself.  First, I'd like to examine the apparent effect of having a candidate from a certain state on the vote share of that state.  I have collected data on all Presidential elections since (and including) 1960.  I only focused on the winners of the elections, because, quite frankly, the other guys lost.

With the exception of 2004 and 2008, I used the two immediate preceding and following election cycles as the controls.  An argument could be made that I should have used more data points, but I believe states shift over time in political representation, and the extra years would only add more noise.  Further, with each year is the national vote share of the winning candidate, the home-state result of the winning candidate, and the difference between the two sets of data.  

Finally, for those (like myself) who prefer visual learning to charts with a lot of numbers, I created bar graphs for each election cycle and state.  Below are the results for those thirteen elections:


Year National Result Massachusetts Result Difference
1952 44.33 45.46 1.13
1956 41.97 40.37 -1.6
1960 49.72 60.22 10.5
1964 61.05 76.19 15.14
1968 42.72 63.01 20.29



YearNational ResultTexas ResultDifference
195641.9743.982.01
196049.7250.520.8
196461.0563.322.27
196842.7241.14-1.58
197237.5233.24-4.28















YearNational ResultNew York ResultDifference
196049.5547.27-2.28
196438.4731.31-7.16
196843.4244.30.88
197260.6758.54-2.13
197648.0247.52-0.5
Year National Result New York Result Difference
1964 38.47 31.31 -17.16
1968 43.42 44.3 0.88
1972 60.67 58.54 -2.13
1976 48.02 47.52 -0.5
1980 50.75 46.66 -4.09

Year National Result Georgia Result Difference
1968 42.72 26.75 -15.97
1972 37.52 24.65 -12.87
1976 50.08 66.74 16.66
1980 41.01 55.76 14.75
1984 40.56 39.79 -0.77

Year National Result California Result Difference
1972 60.67 55 -5.67
1976 48.02 49.35 1.33
1980 50.75 52.69 1.94
1984 58.77 57.51 -1.26
1988 53.37 51.13 -2.24

Year National Result California Result Difference
1976 48.02 49.35 1.33
1980 50.75 52.69 1.94
1984 58.77 57.51 -1.26
1988 53.37 51.13 -2.24
1992 37.45 32.61 -4.84

Year National Result Texas Result Difference
1980 50.75 55.28 4.53
1984 58.77 63.61 4.84
1988 53.37 55.95 2.58
1992 37.45 40.56 3.11
1996 40.27 48.76 8.49

Year National Result Arkansas Result Difference
1984 40.56 38.29 -2.27
1988 45.65 42.19 -3.46
1992 43.01 53.21 10.2
1996 49.23 53.74 4.51
2000 48.38 45.86 -2.52

Year National Result Arkansas Result Difference
1988 45.65 42.19 -3.46
1992 43.01 53.21 10.2
1996 49.23 53.74 4.51
2000 48.38 45.86 -2.52
2004 48.27 44.55 -3.72

Year National Result Texas Result Difference
1992 37.45 40.56 3.11
1996 40.27 48.76 8.49
2000 47.87 59.3 11.43
2004 50.73 61.09 10.36
2008 45.6 55.39 9.79

Year National Result Texas Result Difference
1992 37.45 40.56 3.11
1996 40.27 48.76 8.49
2000 47.87 59.3 11.43
2004 50.73 61.09 10.36
2008 45.6 55.39 9.79

Year National Result Illinois Result Difference
1992 43.01 48.58 5.57
1996 49.23 54.32 5.09
2000 48.38 54.6 6.22
2004 48.27 54.82 6.55
2008 52.87 61.85 8.98


Okay, so that's a LOT of data to digest.  There are clearly some years in which a candidate was awarded a home-state advantage, such as in 1976 for Jimmy Carter and 1992 for Bill Clinton.  Also, there are a lot of states where the effect is much more marginal, such as Reagan in 1980 and 1984.  Perhaps a better way to analyze these numbers is to take the average of all the years a candidate is not from the home-state (within the 5 election cycle window) and compare it to the share of the vote the candidate received in his respective election cycle.  Below is such a comparison:

Year (State) State Average Without State Candidate  State Performance With State Candidate Change
1960 (MA) 8.74 10.50 1.76
1964 (TX) -0.76 2.27 3.03
1968 (NY) -3.31 0.88 4.19
1972 (NY) -3.92 -2.13 1.79
1976 (GA) -9.87 16.66 26.53
1980 (CA) -2.19 1.94 4.13
1984 (CA) -1.92 -1.26 0.66
1988 (TX) 5.95 2.58 -3.37
1992 (AR) -2.75 10.20 12.95
1996 (AR) -3.23 4.51 7.74
2000 (TX) 9.14 11.43 2.29
2004 (TX) 9.14 10.36 1.22
2008 (IL) 5.86 8.98 3.12


The chart above seems to confirm that candidates do typically receive some sort of home-state advantage.  The only candidate who faired worse than his state's average was George H. W. Bush in 1988.  My guess is this is likely to be a factor of him having served as Vice President for eight years prior to running for President.  As such, he was less "from" Texas than he once was.  

However, again, these averages taken of vote share before and after an election could be masking less more or less favorable results by state trends.  For example, before John F. Kennedy was elected in 1960, Massachusetts voted fairly consistently with the national average.  However, after Kennedy was President (and was assassinated), Massachusetts has reliably voted strongly in favor of the Democratic candidate.  Other states, such as California, have also had huge shifts since the beginning of this time period.

To look at perhaps a more accurate result of how a home-state will help in real time, I decided to exclude any data from after the election cycle and only focus on how the state voted in the years prior to the election cycle.  Further, as with the other averages, I excluded years in which a candidate from the home-state was on the ballot.  For example, I excluded George H. W. Bush's Texas result from George W. Bush's data average.  Below are the results:

Year (State) State Average Before State Candidate State Performance With State Candidate Change
1960 (MA) -0.23 10.50 10.74
1964 (TX) 1.41 2.27 0.87
1968 (NY) -4.72 0.88 5.60
1972 (NY) -7.16 -2.13 5.03
1976 (GA) -14.42 16.66 31.08
1980 (CA) -2.17 1.94 4.11
1984 (CA) 1.33 -1.26 -2.59
1988 (TX) 4.69 2.58 -2.10
1992 (AR) -2.87 10.20 13.07
1996 (AR) -3.46 4.51 7.97
2000 (TX) 8.49 11.43 2.94
2004 (TX) 8.49 10.36 1.87
2008 (IL) 5.86 8.98 3.12


As you can see, the number of candidates who received at least a marginal boost from their home-state was relatively unchanged.  The only addition to the group who performed worse was Ronald Reagan in 1984.  However, this can be attributed to the fact that California was becoming much more Democratic after Mr. Reagan was elected in that year.  

Taken as a whole, since 1960, a candidate's home-state has performed better than the average of the election cycles surrounding it by 5.08% and has performed better than the average of its preceding election cycles by 6.28%.  Additionally, if we were to only look at the elections since 1988, the numbers would be similar (though slightly less) at 3.99% and 4.48%, respectively.  

This suggests (and I believe fairly accurately) that a candidate will, on average, get somewhere between a 4% and 5% boost in their home-state, and thus, the home-state advantage theory is validated.  

Therefore, I am confused as to why Mr. Romney chose, out of all his possible home-states, to allow Massachusetts to be his home-state.  Massachusetts will vote for President Obama by a very large margin and is therefore out of reach for Romney.  New Hampshire, on the other hand, is currently giving President Obama only a 5.7 point lead, which is hardly insurmountable.  Perhaps more importantly, Obama is only ahead in Michigan by 1.8%, and it offers 16 electoral votes!  

As I have cautioned numerous times, these polls will change as the election approaches, but one thing is very clear:  Mitt Romney should have chosen a swing state to base his campaign.