I have followed the
last 2 US Presidential Elections pretty closely. It is pretty heady stuff for a
political junkie. The Election happens every 4 years and the endless
discussion, debate and speculation begins almost 3 years in advance. More
importantly, from about 18 months out, there are about 5 polls a week by a
multitude of organizations which makes the whole exercise fascinating, but
somewhat predictable. What particularly struck me was that both in 2008 and in
2012, the eventual winner was predicted by almost everybody and it turned out
to be true.
In the New York Times,
there used to be blog run by Nate Silver which he used to update every day with
the most updated prediction about how many seats the Democratic and the
Republican candidates were going to win and the same split by state. The most
remarkable aspect was that Nate himself did not do any polling. He would simply
review in detail the latest polling that was done by other organizations and
base his prediction on their results, research methodologies, sample selection,
political affiliation and their historical accuracy. For me, it was an
awesome demonstration of the power of analytics.
The results, in
addition to predicting the winner (not very impressive, as discussed above),
Nate actually predicted the results correctly in 49/50 states in 2008 and 50/50
states in 2012! I couldn't help but compare this to what goes on here in India.
Here, the pre- election discourse is dominated by the promiscuity of the
political parties that change partners by the hour. Next will come the
mudslinging and childish populism. Color TV for votes, cows for votes, laptops
for votes, medical insurance for votes, subsidized idli for votes.. its endless
and sickening. Absent from all this will be a single debate on substantive
policy, or ideology for that matter.
Now let’s talk about
psephology, indian style. The political parties will buy news media
organizations and will dictate the results that they would publish. So, for
every contest, state or center, there will be diametrically opposite results
published, irrespective of how close or lopsided the contest really is.
Needless to say, none of them are anywhere close to accurate and the
researchers themselves will proudly admit that they don’t expect to be correct.
Think back to 2004. Nobody thought BJP would lose. In 2009, not one poll
predicted that the Congress would do well. The lazy reasons attributed are
multiparty system, first past the post principle, vote to seat conversion etc.
In my opinion, there
is only one reason. The absence of credible polling and proper analysis. If
there are a few polls done regularly by dispassionate third party
organizations, backed up by professional analysts, we can get to an equally
impressive predictable outcome. Why doesn't it happen here? It’s time somebody
raised the bar.
Where is our Nathan
Selvarathnam???