You are on page 1of 5

Chapter 2 Weighted Voting

A weighted voting system is any voting system in which voters are not equal in
terms of number of votes they control.

We will only consider situations, called motions, where there are only 2 choices,
which will we equate to voting yes or no.

Have 3 elements we need to consider:

1. Players-the voters. If we have n players, we will denote them by P 1, P2, …, Pn.


2. The weights- how many votes each player controls. Denote these by w 1, w2, …
wn. Total number of votes is denoted by V.
3. The quota- Denote by q. The minimum number of yes votes needed for a
proposal to pass. If the quota is less than V/2, it is possible that in a certain vote, a
motion could both pass and not pass. Due to this, we will only consider situations
were 𝑉/2 < 𝑞 ≤ 𝑉 (our quota has to be at least a simple majority).

Denote a generic weighted voting system with n players by [𝑞: 𝑤 , 𝑤 , … , 𝑤 ], where


𝑤 ≥𝑤 ≥⋯≥𝑤 .

Ex. Suppose we have a 4 person (P1,P2,P3,P4) search committee for a job opening. P1
is the head of the committee and so they control 7 votes. P 2 is second in charge and
they control 5 votes. The remaining two people each control 4 votes. Suppose that
a candidate needs 15 votes to move onto the next round of interviews.

Notation for this is:

Situations that can occur


One person-one vote situation: Even though we have a weighted system, the
weights and quota are such that it becomes like every voter is equal.
Dictator: A player whose weight is equal or greater than that of the quota.
Dummies: Players who have no say in the outcome of the voting. If there is a
Dictator, all other players are dummies. However, can still have a dummy with no
dictator.
An important concept that goes with weighted voting is:
Veto Power: When a player is not a dictator, but a motion cannot pass unless this
player votes for it.
This occurs when a player has weight w, with w < q and V-w<q.

For our example, who has veto power?

The above terms demonstrate that just looking at how many votes a player controls
can be misleading. We are going to look at a couple of ways to determine the
“power” a player has.

Important Terms
Coalition: any grouping of one or more players that vote the same way.
Winning Coalition: a coalition that has enough votes to make a motion pass (total is
at least the quota).
Critical Players: Any player in a winning coalition that needs to vote yes for motion
to pass, in other words, with them the coalition is winning and without them it is a
losing coalition.
Winning coalitions for our example, and their critical players:
A player is a dictator if their single-player coalition is a winning coalition.
A player has veto power if they are critical in every winning coalition.
A player is a dummy if they are not critical in any winning coalition.

Banzhaf power Index:


1st way to measure how much voting “power” a player has.

Determine how many times a player is a critical player. We call this number the
critical count for the player. We do this for every player and denote by B 1,B2,…
Then total these up, and call it T. Then determine the ratio B i/T and this is the
Banzhaf power index for Player i. Denote this by 𝛽 .

Banzhaf power distribution:


Collection of Banzhaf power indexes for each player. Sums up to 1 or 100%.

Banzhaf power distribution for our example:


A player with a Banzhaf power index of 1 or 100% is a dictator. A player with a
Bahnzaf power index of 0 is a dummy.
More Terms
Sequential coalition- an ordered list of the players. We use the order to tell us
when each player joins the coalition. With N players, there are
𝑁! = 1𝑥2𝑥3 … 𝑥 𝑁 many sequential coalitions.

Pivotal Player: The player who contributes the first vote that causes a motion to
pass in a sequential coalition. Each winning sequential coalition has one and only
one pivotal player.

Sequential Coalitions and their pivotal players for our example:


Shapley-Shubik Power Index:
2nd way to determine power of a voter.
Using the list of all sequential coalitions determine how many times a player is a
pivotal player and divide that number by 𝑁!. Denote this by 𝜎 , 𝜎 , …

Shapley-Shubik power distribution:


Collection of Shapley-Shubik power index for every player. Sums up to 1 or 100%.

Shapley-Shubik power distribution for our example:

Homework pg. 60 3,5,11,13,17,19,27,29,33

You might also like