Thoughts on the Election: Triumph of Data over Dollars

This post is intended mostly for readers in other countries. Some Americans might find some things that they didn't already know.

Source: Freedom's Lighthouse

When the United States government was created, there were slaves in every state. The nation was rural. Suffrage was restricted to white, male property holders, almost all of whom were Protestants. If those conditions still existed, Mitt Romney would have been elected on Tuesday. As it was, the big cities, women, blacks, Latinos, Jews, Catholics and gays all gave their majorities to Obama. So did young voters. That was why Obama won the popular vote. The white males voted in their majority for Romney.

The changes in the electorate seem to favor the Democrats in future elections, in that the "minorities" are growing as a percentage of the total population, and urbanization continues. Unless the Republicans manage to attract voters from these groups in the future, the Democrats will be increasingly successful in elections.

The U.S. President is chosen by the electoral college, not the popular vote. The system was the result of a compromise between the large and small states in the Constitutional Convention. Thus each state has as many votes in the electoral college as it has members of Congress (i.e. two senators and as many members of the House of Representatives as its population justifies). Four presidents have been elected by the electoral college even though they lost in the popular vote. All but two states have enacted laws specifying that their electoral votes be cast in a block.

Over the history of the United States, granting statehood has been an intensely political process. Before the Civil War, the compromise was to admit free and slave states in equal number -- a compromise made to protect the institution of slavery. During the Civil War the Republican dominated Congress admitted two free states. In the later 19th century, when the Republicans enjoyed the electoral spoils of the Civil War, a number of western states were admitted to the Union, all of which were and are rural with relatively small populations. Currently this fact seems to suggest that Republicans can more easily win the election while losing the popular vote, as did George Bush in 2000.

Before the passage of the Civil Rights Act in the 1960s, the South was solidly Democratic. At the time the Republican party had a conservative wing and a more centrist wing. The Democratic party had a northern liberal wing and a southern conservative wing. With the Nixon Southern strategy, the South became a Republican stronghold, joining a swath of mid western rural states. Those are the red states in the above map. The northeastern and western states are the Democratic stronghold. In the current election there were only a few swing states which were thought to be available to each party (in the middle in the following table).

(The Congress is different. Senators are elected from their states as a whole, but Representatives are elected from their specific electoral districts; there has been a great deal of gerrymandering, so that many congressional districts are safe for one or the other party. This, combined with the advantage enjoyed by incumbents, helps explain why the House of Representatives can have an elected majority belonging to one party while the president elected at the same time belongs to the other party.)

Source: Nate Silver's 538 Blog

The Obama campaign developed, organized and analysed information with exceptional skill in this election. It did so to raise huge amounts of money, to get out the electorate, and to target its efforts to key swing states.

Money

The presidential election of 2012 was the most expensive ever. Total spending in the 2012 election has been estimated at nearly $6 billion. The New York Times estimate of the money raised through September by the presidential candidates, the national party committees and the primary “super PACs” whose sole purpose is to support a candidate was: $934.0 million for Obama and $881.8 million for Romney. Other funding surely was raised and spent by super PACs and issue oriented non-profits.

I think of the Republicans as the party of the more affluent, Democrats as the less affluent. In this election, Obama campaigned on a promise to raise tax rates for high income people, Romney to lower their taxes. Moreover, a Supreme Court decision had declared that corporations had the right to fund political campaigns, and one expects that they would have supported the Republican Romney who ran on his business background and skills. Consequently I thought a year ago that it was likely that the Republicans would raise more money for this presidential election that would the Democrats,

While the Republican party has often outspent the Democratic party in national elections, the Democrats have outspent the Republicans in the last three presidential elections. It seems that the Obama campaign pioneered in 2008 and further developed in 2012 "big data" techniques that were exceptionally successful in identifying potential donors, and exceptionally successful in targeting advertisements and donation requests, as well as online procedures that facilitated small donations. Of course, phone banks were still used as were fund raising events, but the fund raisers were provided with more and better information on who to approach and how to do so. (Read more.)

Getting out the vote

The 2008 election was held when President Bush was unpopular, the United States was entering the Great Recession, and we were involved in two unpopular wars. In 2012, President Obama was relatively unpopular because of the continuing recession and the continuing high levels of unemployment. President Obama won the 2008 election by a much greater margin over Senator McCain than he did the 2012 election over Mitt Romney. Much of the difference must be attributed to the changed conditions that were more favorable for the Democrats in 2008. Some must be attributed to the success of the Romney campaign in making the case for Romney, energizing its base, and getting out the Republican voters.

Still, it seems to be agreed that the Democrats did an exceptional job of getting out the likely Obama voters. This seems to be true in spite of reports of Republican efforts in a number of states to make voting more difficult. The success of the Obama campaign has been attributed in part to developing strong local campaign presence, but also to the big data approach which was effective in helping to identify potential campaign workers, and in identifying the likely Obama voters, not to mention developing effective advertisements to encourage people to vote.

Source: Jedreport.com

Targeting Swing States

It seems clear that the Obama campaign very early on recognized that most states were going to be safely in the Republican or the Democratic camp, and that there would be nine swing states whose electoral votes would determine who won the election. Both campaigns in fact concentrated on these states, but Obama won most of them, notably Florida with 29 electoral college votes, Ohio with 18, and Virginia with 13. Since Romney would have won the election had he obtained 65 more electoral college votes, the 60 votes from these three states alone were almost conclusive. However, Obama also took Pennsylvania, Wisconsin, Iowa, Colorado and Nevada.

Again, the Obama campaign used big data effectively. It had polling data on something like 28,000 potential voters in Ohio alone. It also effectively built local capacity in these states, mobilized its bases there, increased registration, drew in workers from other states, and got out the likely Obama voters.

The techie staff on the Obama 2012 campaign drew on the experience of the 2008 campaign, but was several times larger. No doubt drawing on the advances in management of big data that have occurred in the last four years, it managed to create an integrated information system for the nationwide party use. Using that data it helped the campaign raise more money than the opponents, get out more voters for Obama, and more effectively target swing states.

Nate Silver and the 538 Blog

Nate Silver apparently first came to the attention of a large part of the public by using statistics on professional baseball. He came to my attention by producing very accurate odds for the 2008 national election, especially calling the presidential race very accurately. He triumphed in this election, calling the presidential odds state by state very accurately, predicting the popular vote and the electoral college outcome, and doing exceptionally well in predicting the Senate and House of Representatives outcomes. He tracked the changing odds for months on his 538 blog on the New York Times website.

Silver takes advantage of the large number of surveys done by others, adding their published data to his own model. He weights the data from the various surveys, giving more weight to those with methods he approves of, and more to those that have been successful in predictions in the past. He claims, and I believe, that this weighted combination of many surveys is likely to be more accurate than single surveys.

The 538 blog showed Obama consistently ahead in the popular and electoral votes from June on, and also as the favorite to be elected. He gave increasingly good odds on Obama winning from mid October. During most of this time, the television news broadcasts were saying that the race was very close. By Monday, Silver was citing 10 to 1 odds on an Obama victory. He of course was dead on right. (Might the networks have been trying to maintain interest in their election night broadcasts or their nightly news, or did they simply not have the same quality of analysis that Silver used?)

Conservative pundits and Fox News were predicting a Romney win, and some a Romney landslide. They were very wrong! (Read this, this and this.) It is believed that people like to vote for a winner, and that a general consensus that one candidate will win increases the margin of that candidate's victory. So it is possible that these pundits were deliberately giving false statements. It is also possible that they believed their own biased surveys and that they believed each other. Whatever the reason for their errors, the conservative pundits were frequently very insulting in their public statements about Nate Silver. However, Silver triumphed in the accuracy of his final predictions, giving some credibility to the accuracy of his tracking over the last five months.

Thus the big data team in the Obama campaign triumphed in the support that they provided to the successful campaign. Remember that never has an incumbent president been reelected with so high an unemployment rate and an economy in such perilous condition since Franklin Roosevelt's time -- the election was not an easy win. So too has Nate Silver's serious analysis based on data from many surveys succeeded in providing the public with accurate odds on the results of the election.

At least three individuals are cited as having donated $10 million each to the Romney campaign. Karl Rove, through his two political outfits, American Crossroads and Crossroads Grassroots Policy Strategies, backed unsuccessful Republican presidential contender Mitt Romney with $127 million on more than 82,000 television spots. Big money may well have been trumped by big data and effective analysis.

President Obama Thanks His Campaign Staff