Shapley-Shubik Indices by Modified Multilinear Approximation

Program ssmmle

(Go straight to data input screen.)

This page enables you calculate the Shapley-Shubik power indices for a large weighted voting body of any size, for example the analysis of power in a shareholder voting game for a public corporation with  many hundreds of shareholders, or a large international organisation such as the IMF or World Bank. It uses Leech's modification of the approximation method of multilinear extensions of Guillermo Owen. It can be applied to voting bodies of any size both in terms of number of players and in terms of votes. The voting weights do not have to be integers.

The method uses a combination of direct enumeration and approximation based on a probabilistic voting model. It is necessary to divide the players into "major" and "minor". The "major" players are a small number including the largest and the "minor" ones are the remainder. The algorithm trades off speed against accuracy: the larger is the number of "major" players assumed the slower and more accurate the algorithm. In practice, in most applications, the number of major players need not be greater than 10. The players must be arranged in decreasing order of the size of their respective weights.

References: Shapley and Shubik (1954), Leech (2003), Owen (1972, 1995).



Data Input for ssmmle

Enter your data in the boxes below.

(The numbers are an example which can be overwritten.)


Number of "Major" Players:
       Number of "Minor" Players:


Quota:                                       

Weights
(Enter the values of the weights in decreasing size order separated by spaces.)



Algorithms for Computing Power Indices Home Page