Hybridization, Indian Style

Andrew Muhlestein

Political Science 150

Professor Hawkins

2 November 2012

Blog 7: Duverger's Law

Duverger's law claims that the type of electoral rules, be it the Single Member District Plurality, Proportional Representation, or another system, determines the number of effective parties. The theory goes that SMDP produces two effective parties while PR produces more.

Let's put this theory to the test by comparing it with a sample country, in this case, India, using information from the Inter-Parliamentary Union (http://www.ipu.org/parline-e/reports/2145_B.htm) and some election results from Wikipedia. India uses a plurality system, or a system in which many parties run and the party with the most votes wins the seat, not necessarily (or even usually) with the majority of the votes. This makes it a Single Member District Plurality system, where the country is split into 543 of these constituencies, each of which sends a single representative to Parliament. If the theory is correct, we would expect there to be two effective parties, just as in the United States, that uses a roughly similar system, has two effective parties. While this seems pretty clear-cut, there is a wrinkle if we choose to include the upper house of Parliament, which is elected by members of the lower house using the Proportional Representation method using a single transferable vote, with 12 seats held in reserve for appointment. Parliament also holds two seats in reserve for appointed representatives of native groups. Now let's take a closer look at the numbers.

The district magnitude, or the number of seats assigned to each district, is one. This makes the formula for the election process very simple – within each district each party selects one person who will be the representative. If that party wins, that individual becomes the new seat in Parliament. Because there is only a single seat per district, there is no need to limit the number of parties that can run. This means that there is no threshold, or magic number or percentage of votes, that a party needs to have to get on the official ballot.

Now, in order to figure out how many effective parties there are, and thereby test the effectiveness of the theory, we shall use the formula Neff = 1/Σ(p^2). However, in all 8,070 candidates, representing 1,000 parties stood for election the last election, which makes calculating the number of effective parties more difficult. In order to be reasonable about this, we shall first eliminate all parties that did not receive any seats. Here is resulting data.

Round no 1
Political Group	Total
Indian National Congress (INC)	206
Bharatiya Janata Party (BJP)	116
Samajwadi Party	22
Majority Society Party (BSP)	21
Janata Dal (United)	20
Congrès des "racines" de tous les Indiens	19
Dravida Munnetra Kazhagam (DMK)	18
Communist Party of India (Marxist)	16
Biju Janata Dal	14
Shivsena	11
Nationalist Congress Party	9
All India Anna Dravida Munnetra Kazhagam	9
Independents	9
Telugu Desam Party (TDP)	6
Rashtriya Lok Dal	5
Shiromani Akali Dal (SAD)	4
Rashtriya Janata Dal	4
Communist Party of India (CPI)	4
Janata Dal (Secular)	3
Jammu & Kashmir National Conference (J&KNC)	3
Muslim League Kerala State Committee	2
Revolutionary Socialist Party	2
All India Forward Bloc	2
Telangana Rashtra Samithi	2
Jharkhand Mukti Morcha	2
Bahujan Samaj Party	1
Asom Gana Parishad (Assam People's Federation)	1
All India Majlis-E-Ittehadul Muslimeen	1
Assam United Democratic Front	1
Bahujan Vikas Aaghadi	1
Marumalarchi Dravida Munnetra Kazhagam	1
Bodoland People's Front	1
Haryana Janhit Congress (BL)	1
Jharkhand Vikas Morcha (Prajatantrik)	1
Nagaland People's Front	1
Sikkim Democratic Front	1
Swabhimani Paksha	1
Viduthalai Chiruthaigal Katch	1
Kerala Congress (M)	1

Next we'll eliminate all parties that had less than 10 of the 543 seats. This still leaves us with 10 separate parties. This is the resulting equation.

Neff = 1/ (206/543)^2 + (116/543)^2 + (22/543)^2 + (21/543)^2 + (20/543)^2 + (19/543)^2 + (18/543)^2 + (16/543)^2 + (14/543)^2 + (11/543)^2).

This comes to 1/(0.14392451729529 + 0.04051565377532 + .04564 + 0.00149568084002 + 0.00135662661227 + 0.00122435551757 + 0.00109886755594 + 0.00086824103185 + 0.00066474704001 + 0.00041037955021), which added together is 1/0.23719906920621.

The result is 4.21586814546328, or more reasonably, 4 effective parties. At first this appears to be quite different from our prediction of two parties. Why might that be? The answer lies in the sheer size of India and its very large number of districts. Within most of those districts, there are only two effective parties. But because India is so vast, which of the many parties are the two effective parties varies. In essence, it's as if each district was a microcosm of an entire democracy. An example may make this clearer. Imagine that, all else being equal, somehow every former British colony remained part of the empire. In the UK there are two effective parties, in the US there are two effective parties, in Australia there are two effective parties, etc., exactly as we would predict. However, in each of these theoretical districts, which are the two effective parties is different. So while the theory is not upheld, precisely, it can at least explain away the objectionable result.

Sources
"IPU PARLINE Database: Lok Sabha (House of the People)." IPU PARLINE Database: INDIA (Lok Sabha). Inter-Parliamentary Union, 2009. Web. 2 Nov. 2012. <http://www.ipu.org/parline- e/reports/2145_E.htm>



Wikipedia Contributors. “Elections in India.” Wikipedia, The Free Encyclopedia. 1 Nov 2012. Web. 2 Nov 2012. <http://en.wikipedia.org/wiki/Elections_in_India>

Wikipedia Contributors. “Table of voting systems by country.” Wikipedia, The Free Encyclopedia. 11 Oct 2012. Web. 2 Nov 2012. <http://en.wikipedia.org/wiki/Table_of_voting_systems-

_by_country>