A Caution to hyping ‘Bellwether’ Constituencies as Predictor
Saturday, February 11, 2017
This election cycle
we have seen Prannoy Roy, NDTV hype Bassi Pathana in Malwa District, Punjab and
Shivam Vij, HuffPo focus on Kasganj, Western Uttar Pradesh Bellwether characteristics.
Somehow the impression was created that by just surveying a single bellweather,
it is possible to determine the outcome of an entire state. This blog received
a lot of feedback enquiring whether this holds true which is why this post.
We do believe it is
highly likely that Bassi Pathana in Malwa District, Punjab will vote AAP being
a bastion of AAP and Kasganj in Western UP will similarly vote BJP because
caste dynamics makes it a BJP bastion. Does it mean that as a state Punjab will
vote AAP and UP the BJP? If it does it would be a mere coincidence or chance but
the probability of them doing so this election cycle is very very low as the character of these elections at the state level have changed radically and show strong divergence to those at the constituency level.
Vivek Kaul, Shivam
Vij’s colleague in Huffpo in an incisive article in HuffPo explains why we
should refrain from simplistic deductions and applications of the bellwether
concept. We reproduce extracts of this article below.
The
question is why has this happened? Why has Kasganj voted for the political
party which has won the Uttar Pradesh elections, in each of the 11 assembly
elections since 1974? Is there some deep reason to this?
As
Vij writes: "There is nothing scientific about the bellwether phenomenon and
Kasganj could well get it right this time."I wish he had stopped here. But he goes on
to explain: "Yet, a coincidence lasting 11 election
cycles must be taken seriously. Perhaps it is the constituency's social
composition and geographic location that make it a bellwether. Upper castes,
Muslims, Dalits, Yadavs and other OBCs are all found in this seat. While Lodhs
are the largest community by numbers, that is representative of how non-Yadav
OBCs are the largest caste block, even if a fragmented one, across the
state."
The
point that Vij is trying to make is that Kasganj is like a mini-version of
Uttar Pradesh. And hence has voted for the winning political party since 1974.
Sometimes
things happen and then we go looking for reasons behind them, but that does not
mean that those things happened because of the reasons we came up with. Allow
me to explain this through a few examples.
Paulos' Theory
The mathematician John Allen Paulos talks about
some stocks, mutual funds or analysts doing well over a period of time, merely
by chance. As he writes in A Mathematician Plays the Stock Market: "A
different argument points to the near certainty of some stocks, funds, or
analysts doing well over an extended period merely by chance. Of 1,000 stocks
(or funds or analysts), for example, roughly 500 might be expected to
outperform the market next year simply by chance, say by the flipping of a
coin."
If we were to start with 1974 as the first
iteration, many constituencies in Uttar Pradesh would have voted for the
winning party in that assembly election. In the next assembly election in 1977,
which is iteration 2, the number of constituencies voting for the winning party
in the two assembly elections, would have come down. In the next assembly
election in 1980, which is iteration 3, the number of constituencies voting for
the winning party in the three assembly elections, would have come down even
further.
Survivorship Bias
And so, the iterations worked 11 times, and by the
time 2012 assembly election happened, only one constituency remained which had
voted for the winning party in all the 11 assembly elections in Uttar Pradesh
between 1974 and 2012.
What does this mean in the context of Kasganj? Just
that the constituency has voted for the winning political party in the last 11
assembly elections, it doesn't mean that it will do so the 12th time around as
well. The probability of that happening is very low. And given that there is no
reason that we should be drawing any lessons from it.
Read full article: HERE
0 comments