You are on page 1of 71

APPORTIONMENT

AND
VOTING
Course Intended Learning Outcomes
At the end of the unit, students will be able to:
1. demonstrate an understanding of different methods
used for apportionment and voting; and
2. apply the concepts of apportionment and voting in
real-life situations.
Apportionment is a method of distributing a number of items
proportionally into several groups based on population sizes.
Basic Elements of Apportionment Problems:

State – is a term used to describe the parties having a stake in the apportionment.
Seat – is a set of k identical, indivisible objects that are being divided among n states.
Population – is a set of n positive numbers that are used as the basis for the apportionment of the
seats to the states.
Apportionment problem – is to find a method for rounding standard quota into whole numbers
so that the sum of the numbers is the total number of allocated items.
Quota Rule – is an apportionment method that apportion to a state/ group has a fractional part,
either the integer immediately above, or the integer part of, that state’s/group’s standard quota.
Lower quota – is the standard quota rounded down to the nearest whole number.
Upper quota – is the standard quota rounded up to the nearest whole number.
Standard divisor – is the ratio of population to seats and it is found by dividing the total
population under consideration by the number of seats.
Types of Monotonicity
House monotone – is an apportionment method if the state/group can lose a seat when the size of the total
number of allocation increases.
Population monotone – is n apportionment method when no state/group can lose a seat when only the
population increases.
Quota monotone – is an apportionment method when no state/group can lose a seat whenever its quota
increases.
The Hamilton Method of apportionment was proposed by Alexander Hamilton to
assign voting seats in the House of Representatives to each represented state. This
method is based on the standard divisor and standard quota of the population. The
standard divisor is defined by the formula
SD = P/A
Where P is the total population and A is the total allocation or number of available
seats to be assigned. On the other hand, standard quota is given by:

SQ = G/SD
Where G is the size of the group. The whole number portion of the standard
quota is called lower quota (LQ).
To get the lower quota of a group, round down the standard to a whole
number. The standard quota of a group with the highest decimal part is only
rounded up to a whole number if the sum of the lower quotas is less than the
value of A. The process is repeated until the value of A is reached.
Example 1:
Consider a country with 6 states and 40 seats in the House of Representatives with
population distributed as follows.
Population

STATE A 58,805

STATE B 32,780

STATE C 109,111

STATE D 78,900

STATE E 67,430

STATE F 44,862

TOTAL 391,888
Determine the apportionment of 40 seats to the states using Hamilton Method
Solution:
391,888
SD = = 9797 .2
40
STATE A STATE B STATE C STATE D STATE E STATE F TOTAL
G
SQ =
SD

P 58,805 32,780 109,111 78,900 67,430 44,862 391,888

SQ = G/SD 6.00 3.35 11.14 8.05 6.88 4.58

Lower quota 6 3 11 8 7 5 38
Final Apportionment

6 3 11 8 7 5 40

A 40
SD 9,797.20
Jefferson Method

The Jefferson method of apportionment uses a modified standard divisor (MSD)


which is chosen by trial and error until the sum of the lower quota is equal to the
required number of allocations. The lower quota is the final apportionment
obtained at the final value of the modified divisor. In this method, the modified
standard divisor is smaller than the original standard divisor.
Example 2:

Solve Example 1 using the Jefferson method.

STATE A STATE B STATE C STATE D STATE E STATE F TOTAL

P 58,805 32,780 109,111 78,900 67,430 44,862 391,888

6.53 3.64 12.12 8.76 7.49 4.98


MSQ = P / MSD
Lower quota 6 3 12 8 7 4 40

Final Apportionment

6 3 12 8 7 4 40

A 40

Modified Std

Divisor (MSD) 9,000


The Hamilton and Jefferson methods may give different final apportionments as
shown below.

Apportionment State State State State State State Total

Method A B C D E F

Final Hamilton Apportionment

6 3 11 8 7 5 40

Final Jefferson Apportionment

6 3 12 8 7 4 40
The Webster Method

The Webster method of apportionment is quite similar to the Jefferson


method. Unlike the other method discussed, this method uses ordinary
rounding, making it completely unbiased. The modified standard divisor
is determined using trial and error.
Example 3:

Solve Example 1 using Webster method

Solution:
This time, choose 9,500 as the modified standard divisor. The sum of the lower quotas will be equal
to 40.
STATE A STATE B STATE C STATE D STATE E STATE F TOTAL

P 58,805 32,780 109,111 78,900 67,430 44,862 391,888

6.19 3.45 11.49 8.31 7.10 4.72


SQ = P / MSD

Lower quota 6 3 11 8 7 5 40

Final Apportionment

6 3 11 8 7 5 40

A 40

Modified Std

Divisor (MSD) 9,500


The Huntington-Hill Method

The Huntington-Hill method of apportionment uses the geometric means to round


standard quotas until the sum of the modified quotas equals the exact number of
seats to be apportioned. You may need to modify the standard divisor to get the
exact total allocation. To find the geometric mean, multiply the upper and lower
quotas then get the square root. If the geometric mean is less than the standard
quota, the final apportionment is based on the lower quota. If otherwise, assign the
upper quota.
Example 4:
Solve Example 1 using the Huntington-Hill method
Solution:
Use 9,500 as the modified standard divisor.
STATE A STATE B STATE C STATE D STATE E STATE F TOTAL
P 58,805 32,780 109,111 78,900 67,430 44,862 391,888

SQ = P / MSD 6.19 3.45 11.49 8.31 7.10 4.72


Upper Quota 7 4 12 9 8 5
Lower Quota 6 3 11 8 7 5 40
Geometric Mean 6.48 3.46 11.49 8.49 7.48 5
Final Apportionment

6 3 11 8 7 5 40
A 40
Modified Std

Divisor (MSD) 9,500


Huntington-Hill Number

When there is a need to add one item to one of the several groups in a
population, the additional one item is given to the group with the highest
Huntington-Hill number.

The formula is as follows

G2
H =
A( A + 1)

Where G is the size of the group and A is the current number of items assigned to
this group.
Example 5:
The following table shows the number of computers assigned to the five departments
of a university in Manila and the number of students enrolled in each department. To
which department should a new computer be assigned using the Huntington-Hill
method?

Department Number of Students Number of Computers

1 Chemical Engineering 289 40

2 Mechanical Engineering 785 74

3 Computer Engineering 590 59

4 Civil Engineering 892 85

5 Architecture 345 45
Solution:
Using the formula for the Huntington-Hill number, we obtain the following:
Fairness Criterion and Apportion Paradoxes

The final apportionment using the Hamilton method always yields to either equal to
the lower quota or higher than the lower quota. The said property is one criterion of
fairness for apportionment methods also known as the quota criterion. Thus, the Hamilton
method always satisfies the quota criterion. Although the Hamilton method satisfies the
quota criterion, it may sometimes result in apportion paradoxes when there is a change in
the size of the group, the required number of allocation, and the number of groups. For
example, a change in the population of a group may result in losing one or more items to a
group with relatively lower population. An increase in the number of items to be
apportioned may result in a lower apportionment to some groups even though the
population of the groups are not changed.
To eliminate paradoxes that can occur using the Hamilton method, other methods
were introduced. The three most important are Jefferson, Webster, and
Huntington-Hill methods which, however, may violate the quota criterion.
Unfortunately, all apportionment methods are not perfect. They actually have
some flaws which were proved by Michael Balinski and Peyton Young.
Paradoxes in Hamilton Method

Paradoxical outcomes may exist using the Hamilton method. A new


apportionment may occur to one of the groups which is lower than its original
apportionment when you change the sizes of the groups, the number of the
groups, or the number of items to be allocated. Note the paradoxes cannot occur
in Jefferson, Webster, and Huntington-Hill methods.
Example 6:

A country decides to apportion 200 representatives to three states.

State Population
A 940
B 9,000
C 10,000

a. Use the Hamilton method to find the apportionment for the three states.
b. If one particular state requests 201 seats, what is the final apportionment?
Where will the extra seat go?
Solution:
a. Using the Hamilton method, we obtain the following results.

State State State

A B C Total

State Population 940 9,000 10,000 19,940

Standard Quota 9.43 90.27 100.30

Lower Quota 9 90 100 199

Final Apportionment 10 90 100 200

Allocation Number 200

Standard Divisor 99.7


b. If the number of representatives is changed to 201, the new apportionment will
be as follows:

State State State

A B C Total

State Population 940 9,000 10,000 19,940

Standard Quota 9.48 90.23 100.81

Lower Quota 9 90 100 199

Final Apportionment 9 91 101 201

Allocation Number 201

Standard Divisor 99.2


The Balinski-Young Impossibility Theorem

The Balinski-Young Impossibility Theorem states that it is mathematically


impossible to develop an apportionment method that can avoid all types of
paradoxes and at the same time can satisfy the quota criterion.

Fairness Measurement
The average constituency is used to measure the fairness of an apportionment. It
is defined as follows:

group size
Average Constituency =
number of representative assigned in the group
A group is poorly represented if it has a greater average constituency as compared
to other groups. If a group has a greater average constituency than another group, then
the difference between their average constituencies is called absolute unfairness of the
apportionment. If the absolute unfairness is zero, then the two groups are equally well
represented.

Example 7:

An art guild is having a show and it has decided that the group of painters with 152
members will have 8 booths and the group of sculptors with 95 members will have 5 booths.
Calculate the absolute unfairness of this assignment of booths.
Solution:

The average constituency of the painters is

And the average constituency of the sculptors is

95
= 19
5

The absolute unfairness is 19 – 19 = 0. Thus, the booths are equally well assigned to the two
groups
THE VOTING SYSTEMS
Voting is a tool used by groups of people in making a collective decision. It can be
presented conveniently in terms of an election system in which one can select a particular
candidate out of a set of candidates on the basis of ballots cast by voters

The Majority System and Plurality of Voting


The majority system is the most common voting system applied to an election with
only two candidates. The winner in the majority system requires more than half of the
people voting for an issue or a candidate. Majority means more than 50% of the total
voters or all the other votes combined. Thus, in an election process consisting of two
candidates and 100 voters, the candidate with 51 votes or higher wins the election.
For more than two candidates, it is possible that the one with the most votes
may not constitute a majority. In this case, the winner of this election will be based
on the plurality method of voting. The winner by this method is the candidate with
the most votes or most first-place votes for ranked candidates. In the plurality
system, alternative choices are not considered.
Example 1:
The table below shows the popular vote for three candidates in a student council
election.
Candidate Votes

A 206

B 116

C 92

Total 414
a. Which candidate receives the majority of the votes?
b. Who wins the election?

Solution:
a. The majority requires at least votes. Hence, no one gets the majority
vote.
b. Based on the plurality method of voting, candidate A receives the most votes so he
should be declared as the winner of this election.
Example 2:
The results of an election with 4 candidates and 50 voters are shown in the preference
list below. Using the plurality method of voting, which candidate wins this election?

Candidate Ranking

A 1 4 2 4 2

B 2 3 1 3 4

C 3 1 4 1 3

D 4 2 3 2 1

No. of Votes 12 8 16 10 4
Solution:

Getting all the total of first-place votes, you will obtain the following:

Candidate Ranking Total First-Place Votes

A 1 4 2 4 2 12

B 2 3 1 3 4 16

C 3 1 4 1 3 18

D 4 2 3 2 1 4

No. of Votes 12 8 16 10 4 50

Candidate C has the most number of first-place votes but he wins in this election
according to the plurality method. [Candidate C did not get the majority of the votes (at least
26 votes).]
Borda Count Method

In this method, each candidate is assigned a weight according to their rank in the preference
list. For n number of candidates, the highest rank (the most favourite) will have n points and
the lowest rank (the least favourite) will have 1 point. The Borda count is the sum of these
weights given to a candidate. The candidate who receives the highest Borda count will be
declared the winner.

Example 3:
Using the Borda count method, determine the winner in the preference list in Example 2.
Solution:
n=4 CANDIDATE A n=4 CANDIDATE C
Rank Rank Point No. of Votes Total Rank Rank Point No. of Votes Total
1 4 12 48 2 3 12 36
4 1 8 8 1 4 8 32
2 3 16 48 4 1 16 16
4 1 10 10 1 4 10 40
2 3 4 12 3 2 4 8
Borda Count = 126 Borda Count = 132

n=4 CANDIDATE B n=4 CANDIDATE D


Rank Rank Point No. of Votes Total Rank Rank Point No. of Votes Total
3 2 12 24 4 1 12 12
3 2 8 16 2 3 8 24
1 4 16 64 3 2 16 32
3 2 10 20 2 3 10 30
4 1 4 4 1 4 4 6
Borda Count = 128 Borda Count = 114

The candidate with the highest Borda count is candidate C. Thus, candidate C wins the election.
Plurality with Elimination

Plurality with elimination is a variation of the plurality method in which the


alternative choices of the voters are taken into consideration.

The candidate with fewest number of first-place votes is first eliminated. In case
there are two alternatives with the same lowest votes, then both are to be eliminated. The
remaining candidates are re-ranked with the assumption that voters’ preference do not
change from round to round.
Example 4
Answer example 2 using the method of plurality with elimination

Candidate Ranking

A 1 4 2 4 2

B 3 3 1 3 4

C 2 1 4 1 3

D 4 2 3 2 1

No. of Votes 12 8 16 10 4

Solution:
For round 1, candidate D should be eliminated.
Candidate Total First-Place Votes

A 12

B 16

C 28

D 4
Round 2
Candidate Ranking

A 1 4 2 4 2

B 3 3 1 3 4

C 2 1 4 1 3

No. of Votes 12 8 16 10 4

Candidate Total First-Place Votes

A 12

B 16

C 8 + 10 = 18

Candidates A and B have the same lowest first-place votes, so both should be eliminated. Thus,
the winner for this method is candidate C.
The Top Two Runoff Method

The candidates with the most number of first-place votes are removed from the
preference list and are then re-ranked for new preference list. The one with the
higher first-place votes in the new preference list between these two candidates will
be declared the winner. The top two runoff method satisfies the majority criterion.
Example 5

Apply the top two runoff method to the preference list from Example 2.

Solution:

Counting the total first-place votes for each candidate, you will obtain the following.

Candidate Ranking Total First-Place Votes

A 1 4 2 4 2 12

B 2 3 1 3 3 16

C 3 1 4 1 4 18

D 4 2 3 2 1 4

No. of Votes 12 8 16 10 4

The top two candidates with the most number of first-place votes are candidates B and C.
Thus, remove candidates A and D from the list. This gives us the following preference list.
Candidate Ranking Total First-Place Votes

B 1 2 1 2 1 32

C 2 1 2 1 2 18

No. of Votes 12 8 16 10 4

Thus, candidate B wins the election. Note that in the plurality method, candidate C wins the
election.
Approval Voting System

In this type of voting, a voter may choose more than one option or candidate. Each vote
coming from one voter will be counted as one vote, and the one with the most number of
total votes will be declared the winner. For large elections, approval voting may be
considered a better method because it measures the overall support for a candidate.
Example 6:

The members of a scholarship council have picked their choices from the top four applicants for a
scholarship. The results are indicated in the table of preference below.
Choices Number of Votes

Harry, Liam, and Louis 5

Harry, Liam, and Niall 3

Niall and Louis 2

Harry and Niall 4

Who got the scholarship using the approval voting system?


Solution:

Tallying the votes, you will obtain the following:

Candidate Number of Votes

Harry 5 + 3 + 4 = 12

Liam 5+3=8

Louis 5+2=7

Niall 3+2+4=9

Hence, Harry gets the scholarship with a total of 12 votes.


The Method of Pairwise Comparison

In the method of pairwise comparison, each candidate is compared head-to-head


with each of the candidates. The candidate with the most number of wins from these
comparisons will be declared the winner. This method satisfies the majority, Condorcet,
and monotonicity criteria of fairness in voting which will be discussed in detail in the
next section.
Example 7:

The members of a mechanical engineering club consisting of 165 active members were asked
to decide in which sports event the club should participate in the coming intercollegiate
games. If the members marked their ballots as shown in the preference table below, which
event wins as the top choice using the method of pairwise comparison?

Number of Voters 26 60 35 44

1st choice volleyball basketball football volleyball

2nd choice basketball volleyball basketball basketball

3rd choice swimming swimming volleyball football

4th choice football football swimming swimming


Solution:
Number of Voters 26 60 35 44 Total First-Place Votes

Volleyball (V) 1 2 3 1 70

Basketball (B) 2 1 2 2 60

Football (F) 4 4 1 3 35

Swimming (S) 3 3 4 4 0

Possible comparison are highlighted in the table show

V B F S

S
Number of Voters 26 60 35 44 Total First-Place Votes Number of Voters 26 60 35 44 Total First-Place Votes

V(Winner) 1 2 3 1 26 + 44 = 70 B (winner) 2 1 2 2 60

B 2 1 2 2 60 F 4 4 1 3 35

Number of Voters 26 60 35 44 Total First-Place Votes Number of Voters 26 60 35 44 Total First-Place Votes

B (winner) 2 1 2 2 60
V (winner 1 2 3 1 26 + 44 = 70

F 4 4 1 3 35
S 3 3 4 4 0

Number of Voters 26 60 35 44 Total First-Place Votes Number of Voters 26 60 35 44 Total First-Place Votes

V (winner) 1 2 3 1 70 F(Winner) 4 4 1 3 35

S 3 3 4 4 0 S 3 3 4 4 0
Below is the summary of the winner in each comparison

V B F S

V B V V

B B B

F S

Form the above table, basketball has the most number of wins. Therefore, this option
should be the top choice by pairwise comparison method.
Measuring Fairness in Voting

In voting, what does the word fair mean? When are the results of an election or a
particular voting method considered fair? People normally think that a majority candidate is
a very strong candidate and it is not fair when candidate loses an election. This thinking
leads to one notion of fairness which is the majority criterion. On the other hand, a single
example of an election that fails to meet the fairness criterion would be sufficient to show
that a particular method is not fair. The following are the four basic criteria of fairness.
1. The Majority Criterion
If a candidate receives the majority of first-place votes in an election, then this candidate
should be declared the winner. The method of plurality is an extension of the majority
system. If the candidate is a majority candidate then clearly he should also be the plurality
winner. In some cases, the winner may not have the approval of majority; nevertheless, the
plurality method satisfies the majority fairness criterion.
2. The Condorcet Criterion
If a candidate is a Condorcet candidate, then he should be the winner of the election. A
candidate is said to be a Condorcet candidate if he wins in all of the one-to-one comparison
with other candidates. The pairwise comparison method is the only method satisfies the
Condorcet criterion of fairness.
3. Monotonicity Criterion
Suppose candidate x wins an election. If the supporters of other candidates change
their votes to support candidate x, then this candidate must win the new election as
well. The monotonicity fairness criterion is satisfied by plurality, pairwise comparison,
and Borda count methods.
4. Independent of Irrelevant Alternatives (IIA)
Suppose candidate x wins an election. If a losing candidate withdraws from the
race, then candidate x should remain the winner of the new election.
Arrow’s Impossibility Theorem
It is impossible to develop a voting method that will satisfy all of the four basic
criterion of fairness.
The following table summarizes the performance of the voting method discussed in
this chapter with respect to the four fairness criteria. The term “yes” means the method
always satisfies the criterion and “no” means it may violate the said criterion.

Fairness Criterion Plurality Borda Count Pairwise Comparison Top Two Runoff

Majority Yes No Yes Yes

Condorcet No No Yes No

Monotonicity Yes Yes Yes No

IIA No No No No
THE WEIGHTED VOTING SYSTEMS
A weighted voting system is a biased voting system where voters have different
amounts of influence in the outcome of an election. The best examples are the stockholders
of a company and the United Nations Security Council.
The weighted voting system W is characterized by three elements: n players, the quota
q, and the weights w of voters. Conveniently, it can be expressed as follows.

with

The minimum number of votes needed to approve a resolution or to pass a measure is


called a quota. The weight of a voter is the number of votes controlled by the voter. There
are different types of weighted voting system.
Types of Weighted Voting Systems
1. One Voter – One Vote System
Each person has only one vote. Thus, a majority of the votes is required to pass a
resolution.

Example 1:

Since the total number of votes is 7, then q = 4 is the required number of votes to pass a
resolution.
Hence, the quota is q = 4.
2. Dictatorship
This type of system has one person whose weight is enough to pass any resolution.

Example 2:

In this system, even if the other 4 candidates vote for the resolution, the sum of votes
is only 7 which is less than the quota of 9 votes.

3. Null System
This is a weighted voting system which cannot pass any resolution because the sum
of all votes is always less than the quota.
Example 3:

Hence, and 16 < q which is equal to 17.

4. The Veto Power System


This is one type of weighted voting system where each voter has a veto power. This means
that if one voter does not vote, no resolution will be passed. This type of voting system will
occur when the quota is equal to the sum of all the votes.

If at least one voter in a weighted voting system has veto power, the system is
considered to be a veto power system.
Example 4:
Here, which is equal to the quota, q = 18.

Even the voter with one vote will have a veto power.
Coalition
A coalition in a voting system is an alliance formed by a group of voters with a common goal
which is either to vote in favor or against a resolution. Below are some of the terms related in a
coalition.
1. A winning coalition is a group of voters whose sum of all votes is greater than or equal to the quota.
2. A losing coalition is a group of voters whose sum of all votes is less than the quota.
3. A critical voter is a voter who turns a winning coalition into a losing coalition if he or she leaves for
another group.
4. A blocking coalition is a group of voters who can prevent the approval of a resolution.
5. A dummy is a voter whose weight does not affect any voting outcome. Note that when a system has
a dictator, all other players are considered dummies.
The number of possible coalition with n voters is the number of possible subsets that can be
formed from these voters. This includes the set containing all the voters and the singleton
subsets.
The number of possible coalition with n voters is given by
The United Nations Security Council
The United Nations Security Council (UNSC) is one of the 6 principal organs of the United
Nations which is given an executive power to make resolutions that will maintain peace and
security in the international community. The council members are limited to only 15
countries: 5 permanent members (US, China, France, United Kingdom, and Russia) and 10
non-permanent members. Each of the five members is given a veto power. Thus, if one of
them votes against a resolution of the council, it will fail. The decision of the council also
requires that 9 members must vote in favour of the resolution. Based on the current system,
the UN Security Council is a veto power system.
Example 5:
Express the UNSC as a weighted voting system. Find the quota and the voting weight of each
member.
Solution:
Let q be the quota for a resolution to pass and w be the voting weight of each permanent
member. Assuming a voting weight of 1 vote for each of the non-permanent members, you
have the following inequalities.
Condition 1: If one permanent member votes against a resolution, it fails.
Note that even if all 10 permanent vote for the resolution, if one permanent member will vote
against it, the quota will not be met.
Condition 2: Nine members including the five permanent members must vote for the
resolution.

Combining the two inequalities,

The smallest integer number after 6 is 7. Hence, w = 7, q = 39, and the weighted system of
the UNSC is defined by
Example 6:
The weighted voting system for a company owned by 5 people is given by

a. Find all the winning coalitions.


b. For each winning coalition, determine the critical voters.

Solution:
Note that
Owners Shares
A 425
B 250
C 175
D 350
Quota = 601
The total possible coalition are

The winning coalitions (WC) are the combinations whose total votes are greater than the
quota 601. These are {A, B}, {A, D}, {A, B, C}, {A, B, D}, {B, C, D}, {A, C, D}, AND {A, B, C, D}.
If a critical voter of a winning coalition is removed, the winning coalition becomes a
losing coalition. The following table summarizes the critical voters for each winning
coalition.
Coalition Total Winning Coalition Critical Voters
1 A 425
2 B 250
3 C 175
4 D 350
5 AB 675 YES A&B
6 AC 600
7 AD 775 YES A&D
8 BC 425
9 BD 600
10 CD 525
11 ABC 850 YES A&B
12 ABD 1025 YES A
13 BCD 775 YES B, C & D
14 ACD 950 YES A&D
15 ABCD 1200 YES None
QUOTA = 601 >601
Banzhaf Power Index
The Banzhaf Power Index (BPI) measures the power of a voter in a weighted voting
system. The value of BPI ranges from 0 to 1. If the BPI value of voter A is greater than of voter
B, then A has greater power than B. This means that A can be part of more winning coalitions
than B. The voting power of a voter is measured by the Banzhaf Power Index which is defined
as follows:

the number of times a voter becomes a critical voter


B=
the total number of critical voters in all winning coalitions
Example 7:
Compute the voting power of each voter in Example 6.
Solution:
The total number of critical voters in all 7 winning coalitions is 12.
Owners B

A 5/12 0.42

B 3/12 0.25

C 1/12 0.08

D 3/12 0.25

Owner A has the most power and owner C has the least power.
THANK YOU!!!

You might also like