How Obama won


Sunday, November 11, 2012

Print this page Email A Friend!

BARACK Obama scored a resounding victory in the US elections last Tuesday to secure his second and final term as president of the United States. In an election that at one stage appeared to be really too close to call, it is likely he could win as many as 332 of the 538 Electoral College votes, once the issue of Florida is sorted out.

From all polls immediately leading up to November 6, the verdict was a dead heat and there was speculation that Mitt Romney could very well snatch one or two of the battleground states, in addition to taking back two of those won by Obama against the tide in 2008, a possibility that would tilt the balance in his favour.

Perhaps the fact that several of the polls two or three days before the election showed Obama with a razor-thin lead sobered up the Republicans somewhat and the hectic schedule of trips across the country in those last three days by both Obama and Romney suggested that they themselves fully recognised that those were critical days in certain key states.

I have not done any original polling work in any of the states, but certainly did a very detailed evaluation of the historical voting pattern in the US presidential elections over the last century, from 1860 through to 2008, to better understand how Americans were likely to vote.

Armed with this information, the period of review was narrowed to the last three elections — 2000, 2004 and 2008. These data were matched with results of the more credible polling outfits over the last week of the campaigning that allowed me to come up with my own assessment of the likely outcome of the election.

One very obvious pattern that emerged over these years of evaluation is that, for the better part of the century, there are states that have voted solidly Democrat or solidly Republican, meaning that in over 66 per cent of the cases, the states have stayed with one particular party.

Another clear finding from that evaluation is that the respective states tend to give each party a fairly long run, that is, they have voted for either the Democrats or Republicans for periods of four or more consecutive presidential elections — a factor which feeds into the notion of strong loyalty over protracted periods.

This evaluation, along with detailed and objective scrutiny of the results of polls from the more credible polling units over the last three to five days helped me to form a clear notion of how the 2012 presidential battle between Obama and Romney would go.

It was indeed a surprise to me to see how easily an Obama victory could be possible, especially knowing something of the efficiency of the ground strategy of Obama, having talked with persons who actually worked on the Obama team in Illinois and elsewhere.

The obvious appeal to the immigrant vote, especially recognising the growing importance of the Latino vote and the anticipated strong support of the African-American segment, coupled with overtures to females were key factors in his campaign strategy.

Moreover, he was able to maintain the large majority of his highly successful ground campaign team from 2008, and it appeared that Romney failed to recognise, perhaps very much like our own Labour Party in the 2011 election, that registering and mobilising the voters was a critical component of a campaign of this nature.

On the evening of the election, before the polls closed, I decided I would share my own findings with CVM TV. These were carried on the major news at 8:00 pm. I reiterate that I did no polls in the US and therefore did not have the benefit of original scientific information to guide me, as we do for elections in Jamaica.

For this reason, and in keeping with our professional perspective, I was reluctant to air my own evaluation. What did I find? Based on this historical analysis, the following formed the basis for a projection as to how the elections would go.

Of the 51 states, Obama was certain to win in 21 (see graphic).

These states have voted consistently Democrat and, certainly over the last three elections, have demonstrated a close loyalty to Obama's party. These states alone guaranteed Obama a total of 257 votes, 13 shy of the magic number of 270.

Romney, on the other hand, was certain to win in 22 of the 51 states.

But this would guarantee him no more than 180 of the Electoral College votes. This factor alone in my estimation was sufficient assurance that Obama would win the race fairly comfortably. The other eight states, namely Colorado (9), Florida (29), Indiana (11), Nevada (6), North Carolina (15), Ohio (18), Virginia (13) and Iowa (6), have more often than not voted solidly Republican, but in the 2008 election all voted for Obama.

With some of the opinion polls in some of these states, especially Ohio, favouring Obama, barring any slips in the sure Democratic states shown in the graphic, this alone would have tipped Obama over the 270 required to win the presidency.

In my summation then, of the historical voting pattern, coupled with the opinion poll results, the outcome of the election was never in doubt.

The scenario that I presented to CVM ahead of the closure of any of the polls in the 51 states was as follows, with 43 of the states decided and four of the remaining eight favouring Obama.

"Obama assured of 257 Electoral College votes;

Romney assured of 180 Electoral College votes;

If Obama holds these and wins either Florida or Ohio he is well over the 270;

If Obama fails to win either of these, but wins any two of Colorado (9), Nevada (6) and Indiana (11), he would have secured the presidency."

The fact is, he lost Indiana, one of the states I had earmarked him to win in my calculations above, but secured as replacements, Virginia, Colorado and Nevada; which gave him another 28, so that in addition to Ohio, which was always trending towards Obama, he gained another 46, which pushed him over the 300 mark.

Based on this calculation therefore, I estimated that Obama would win between 295 and 301 votes with Florida hanging in the balance. At present, Obama has 303 electoral votes with Florida still undecided, and if he takes Florida, as is expected, he will end up with 332, very much in line with the projections based on the historical voting pattern, matched with the latest opinion poll data.

The validity of using historical data to project the outcome was strengthened by the fact that the opinion polls showed consistently that Obama was ahead, albeit marginally, and hence there were no overarching negatives that could cause a reversal of fortune in those states that were clearly Democrats.

By extension, where opinion polls showed the incumbent in those dominant states trailing, then this would weaken the ability to use the historical data as a mark of likely voting intention. Hence, this model cannot always be used, but instead would need to be modified based on the trend of opinion polls in each of the states.

Having done this analysis ahead of the closure of the polling stations, it was difficult to see Obama losing this election and the surprise to me then was that most of the official media and most of the opinion polls played this out as a very close election, one too close to call.





1. We welcome reader comments on the top stories of the day. Some comments may be republished on the website or in the newspaper � email addresses will not be published.

2. Please understand that comments are moderated and it is not always possible to publish all that have been submitted. We will, however, try to publish comments that are representative of all received.

3. We ask that comments are civil and free of libellous or hateful material. Also please stick to the topic under discussion.

4. Please do not write in block capitals since this makes your comment hard to read.

5. Please don't use the comments to advertise. However, our advertising department can be more than accommodating if emailed:

6. If readers wish to report offensive comments, suggest a correction or share a story then please email:

7. Lastly, read our Terms and Conditions and Privacy Policy

comments powered by Disqus



Today's Cartoon

Click image to view full size editorial cartoon