And the search for an Indian Nate Silver
Writing in The New York Times, Margaret Sullivan argues,
His [Nate Silver] entire probability-based way of looking at politics ran against the kind of political journalism that The Times specializes in: polling, the horse race, campaign coverage, analysis based on campaign-trail observation, and opinion writing, or “punditry,” as he put it, famously describing it as “fundamentally useless.” Of course, The Times is equally known for its in-depth and investigative reporting on politics.
His approach was to work against the narrative of politics – the “story” – and that made him always interesting to read. For me, both of these approaches have value and can live together just fine. [link]
Nate Silver’s great insight in the 2012 US presidential elections wasn’t that Barack Obama was likely to be reelected president. Anyone who dispassionately read polls reached the same conclusion. Even Realclear politics which merely averages polls performed nearly as well as Nate Silver’s much heralded model. What Silver was really arguing—and what really irrigated the likes of Joe Scarborough—was that contrary to reports of daily swings, the race was consistently in favor of President Obama for pretty much the entire duration of the campaign. Sure, his numbers dropped a little bit after his stumble during the first presidential debate but Obama remained the consistent favorite.
This minimalistic numbers based approach naturally militates against the traditional political reporting where every stumble or mistake is highlighted and the race swings widely from one candidate to another. The kind of traditional political reporting where bar interviews and the body language of advisors are supposed to suggest which way the election is heading. Hold on, said Silver: the basic nature of the race hadn’t changed despite media obsession over each and every surreptitiously recorded YouTube video. In other words, Nate Silver attempted to rob the elections of their traditional excitement.
As the election season arrives in India, this debate may have interesting parallels in India. The Indian media has started to pay attention to data journalism though it remains limited by Western standards. And as media houses release polls on the 2014 elections, there is a yearning for Indian Nate Silver who can help make sense of the often contradictory results. But it faces at least two significant challenges.
First, there is simply not enough polling to yield panel data which is often far more interesting and explanatory than simply looking at numbers. Polling a country like India is terribly expensive and the Indian media has limited resources. Second, polling India is really hard even if the best methodology is employed and adequate resources are devoted. For instance, the US is a two-party country where the vast majority of states are non-competitive at the national level. You don’t need polling or any fancy statistical analysis to understand which way California is likely to vote. In sharp contrast, India has multitudes of regional parties and the national formations are restricted to a minority of states. And to state the obvious, polling Delhi and rural Maharashtra are entirely different. Ultimately, the predictive power of Nate Silver’s model was driven by the quality of polling; if the polls were wrong, so would have been his model.
So the search for Indian Nate Silver is likely to be futile exercise. Traditional political journalism will continue to dominate reporting in India with polls making for some intriguing TV discussions but having limited salience otherwise.
p.s To be clear, this is not to suggest that polls in India are entirely useless. State level polls at least in two-party states usually get it right. However, their significance at the national level—beyond suggesting broad trends—is limited.