MATH 1340 Mathematics & Politics

Similar documents
MATH 1340 Mathematics & Politics

Exercises For DATA AND DECISIONS. Part I Voting

Social welfare functions

MATH 1340 Mathematics & Politics

The search for a perfect voting system. MATH 105: Contemporary Mathematics. University of Louisville. October 31, 2017

SOCIAL CHOICES (Voting Methods) THE PROBLEM. Social Choice and Voting. Terminologies

Desirable properties of social choice procedures. We now outline a number of properties that are desirable for these social choice procedures:

Chapter 9: Social Choice: The Impossible Dream Lesson Plan

Fairness Criteria. Majority Criterion: If a candidate receives a majority of the first place votes, that candidate should win the election.

Chapter 1 Practice Test Questions

12.2 Defects in Voting Methods

Elections with Only 2 Alternatives

MATH4999 Capstone Projects in Mathematics and Economics Topic 3 Voting methods and social choice theory

The Impossibilities of Voting

Voting Criteria April

answers to some of the sample exercises : Public Choice

2-Candidate Voting Method: Majority Rule

Arrow s Impossibility Theorem

Many Social Choice Rules

Voting Lecture 3: 2-Candidate Voting Spring Morgan Schreffler Office: POT Teaching.

Computational Social Choice: Spring 2007

Economics 470 Some Notes on Simple Alternatives to Majority Rule

1.6 Arrow s Impossibility Theorem

Intro to Contemporary Math

Introduction: The Mathematics of Voting

The Mathematics of Voting. The Mathematics of Voting

Mathematics and Social Choice Theory. Topic 4 Voting methods with more than 2 alternatives. 4.1 Social choice procedures

Mathematical Thinking. Chapter 9 Voting Systems

Voting: Issues, Problems, and Systems, Continued

Voting: Issues, Problems, and Systems, Continued. Voting II 1/27

Math for Liberal Studies

Measuring Fairness. Paul Koester () MA 111, Voting Theory September 7, / 25

Fairness Criteria. Review: Election Methods

Write all responses on separate paper. Use complete sentences, charts and diagrams, as appropriate.

Social Choice: The Impossible Dream. Check off these skills when you feel that you have mastered them.

Head-to-Head Winner. To decide if a Head-to-Head winner exists: Every candidate is matched on a one-on-one basis with every other candidate.

9.3 Other Voting Systems for Three or More Candidates

In this lecture we will cover the following voting methods and fairness criterion.

Math for Liberal Arts MAT 110: Chapter 12 Notes

Chapter 10. The Manipulability of Voting Systems. For All Practical Purposes: Effective Teaching. Chapter Briefing

Introduction to Theory of Voting. Chapter 2 of Computational Social Choice by William Zwicker

Today s plan: Section : Plurality with Elimination Method and a second Fairness Criterion: The Monotocity Criterion.

(67686) Mathematical Foundations of AI June 18, Lecture 6

Rationality of Voting and Voting Systems: Lecture II

Recall: Properties of ranking rules. Recall: Properties of ranking rules. Kenneth Arrow. Recall: Properties of ranking rules. Strategically vulnerable

The Manipulability of Voting Systems. Check off these skills when you feel that you have mastered them.

How should we count the votes?

Voting Definitions and Theorems Spring Dr. Martin Montgomery Office: POT 761

Approaches to Voting Systems

Introduction to the Theory of Voting

CSC304 Lecture 16. Voting 3: Axiomatic, Statistical, and Utilitarian Approaches to Voting. CSC304 - Nisarg Shah 1

Voting rules: (Dixit and Skeath, ch 14) Recall parkland provision decision:

Social Choice. CSC304 Lecture 21 November 28, Allan Borodin Adapted from Craig Boutilier s slides

Math Circle Voting Methods Practice. March 31, 2013

Chapter 9: Social Choice: The Impossible Dream Lesson Plan

Section 3: The Borda Count Method. Example 4: Using the preference schedule from Example 3, identify the Borda candidate.

Make the Math Club Great Again! The Mathematics of Democratic Voting

Chapter 9: Social Choice: The Impossible Dream

Voting: Issues, Problems, and Systems. Voting I 1/36

Homework 7 Answers PS 30 November 2013

CS 886: Multiagent Systems. Fall 2016 Kate Larson

The Mathematics of Voting

Voting Criteria: Majority Criterion Condorcet Criterion Monotonicity Criterion Independence of Irrelevant Alternatives Criterion

Notes for Session 7 Basic Voting Theory and Arrow s Theorem

Social Choice Theory. Denis Bouyssou CNRS LAMSADE

Syllabus update: Now keeping best 3 of 4 tests

Voting. Hannu Nurmi. Game Theory and Models of Voting. Public Choice Research Centre and Department of Political Science University of Turku

Voting: Issues, Problems, and Systems. Voting I 1/31

Democratic Rules in Context

Math116Chap1VotingPart2.notebook January 12, Part II. Other Methods of Voting and Other "Fairness Criteria"

VOTING SYSTEMS AND ARROW S THEOREM

Voting Protocols. Introduction. Social choice: preference aggregation Our settings. Voting protocols are examples of social choice mechanisms

The Mathematics of Voting

Computational Social Choice: Spring 2017

Math for Liberal Studies

Algorithms, Games, and Networks February 7, Lecture 8

Explaining the Impossible: Kenneth Arrow s Nobel Prize Winning Theorem on Elections

: It is mathematically impossible for a democratic voting method to satisfy all of the fairness criteria was proven in 1949.

Section Voting Methods. Copyright 2013, 2010, 2007, Pearson, Education, Inc.

Constructing voting paradoxes with logic and symmetry

Is Democracy Possible?

Section Voting Methods. Copyright 2013, 2010, 2007, Pearson, Education, Inc.

Name Date I. Consider the preference schedule in an election with 5 candidates.

The mathematics of voting, power, and sharing Part 1

Voting: Issues, Problems, and Systems

Voting and preference aggregation

Voting and preference aggregation

In deciding upon a winner, there is always one main goal: to reflect the preferences of the people in the most fair way possible.

MULTIPLE CHOICE. Choose the one alternative that best completes the statement or answers the question.

Lecture 12: Topics in Voting Theory

Majority- more than half of the votes Plurality- the most first place votes. The Majority Criterion

Arrow s Conditions and Approval Voting. Which group-ranking method is best?

Social choice theory

Voter Sovereignty and Election Outcomes

The Math of Rational Choice - Math 100 Spring 2015

Public Choice. Slide 1

Simple methods for single winner elections

1 Voting In praise of democracy?

CSC304 Lecture 14. Begin Computational Social Choice: Voting 1: Introduction, Axioms, Rules. CSC304 - Nisarg Shah 1

The Mathematics of Voting and Elections: A Hands-On Approach. Instructor s Manual. Jonathan K. Hodge Grand Valley State University

Transcription:

MATH 1340 Mathematics & Politics Lecture 6 June 29, 2015 Slides prepared by Iian Smythe for MATH 1340, Summer 2015, at Cornell University 1

Basic criteria A social choice function is anonymous if voters exchanging preference ballots doesn t change winners. A social choice function is neutral if switching a winner with another candidate in all preference orders turns the latter into a winner. A social choice function is monotone if moving a winner up one position in a voter s preference order does not cause them to become a loser. A social choice function satisfies the majority criterion if whenever a candidate receives a majority of first-place votes, that candidate is the unique winner. A social choice function is decisive if it only selects unique winners. 2

Basic criteria (cont d) Proposition: Any social choice function that satisfies anonymity and neutrality must violate decisiveness. Proof (from class): Suppose that our social choice function is anonymous and neutral, and we have just 2 candidates, and 2 voters (any even number will work). Consider the following profile: A B B A If our method was decisive, we must have a unique winner, say A. But now, consider the profile in which we switch candidates A and B: B A A B By neutrality, B must be a winner in this second profile. However, this switch had the same effect as the voters exchanging their preference ballots. Thus, anonymity implies that A must still be the winner in this profile. But then there is a tie in this second profile, contradicting that our method was decisive. Thus, it must not be decisive. 3

Basic criteria (cont d) Proposition: Any social choice function that satisfies anonymity and neutrality must violate decisiveness. Proof (cont d): A similar trick works for any number of candidates. Consider the case where we have 3 candidates, and 6 voters. Consider the profile: A A B B C C B C A C A B C B C A B A Suppose that A was the unique winner. Consider the profile we get from switching A and C: C C B B A A B A C A C B A B A C B C By neutrality, C must become a winner, but this switch had the same effect as voters exchanging ballots (1 with 6, 2 with 5, 3 with 4), so anonymity implies A is still a winner. Similarly if B or C was the unique winner in the original profile. Thus the method cannot be decisive. QED 4

Basic criteria (cont d) Here s where we stand: Anonymous Neutral Monotone Majority Decisive Plurality Yes Yes Yes Yes No Anti-Plur. Yes Yes Yes No No Borda Yes Yes Yes No No Hare Yes Yes No Yes No Coombs Yes Yes No Yes No Copeland Yes Yes Yes Yes No Dictatorship No Yes Yes Yes Yes Monarchy Yes No Yes No Yes All-ties Yes Yes Yes No No 5

The Pareto criterion A social choice function satisfies the Pareto criterion if whenever every voter prefers one candidate over another, say X over Y, the function does not select Y as a winner. A candidate is Pareto optimal if no single other candidate is preferred by all voters. The Pareto criterion demands that only Pareto optimal candidates are permitted to win. Vilfredo Pareto (1848-1923) This term comes from economics: An alternative is Pareto optimal if no single other alternative is unanimously preferred. 6

The Pareto criterion (cont d) For example: Consider the following profile: A B A B C A B A B C C C Since every voter prefers A over C, C cannot win if this method is Pareto. Both A and B are Pareto optimal; no single candidate is preferred by all voters over A, and likewise for B. 7

The Pareto criterion (cont d) However, consider the following profile: A B A B C C B A B A C C C is now Pareto optimal, because neither A nor B are unanimously preferred over C. Being Pareto does not rule out any candidates as winners in this profile. Caution: The Pareto criterion does not say that if a candidate is Pareto optimal then they must win (we ll see that plurality is Pareto, so the above profile would be a counterexample; C is Pareto optimal but is not a winner in the plurality method), only that non-pareto candidates must not win. 8

The Pareto criterion (cont d) A social choice function satisfies the unanimity criterion if whenever a candidate receives every first-place vote, they are selected as the unique winner. If a social choice function satisfies the majority criterion, it must satisfy the unanimity criterion. Why? From class: If a candidate receives every first-place vote, then they ve received a majority of first-place votes. (Exercise 3.5 in R&U) If a social choice function satisfies the Pareto criterion, it must satisfy the unanimity criterion. Why? 9

The Pareto criterion (cont d) (Exercise 3.5 in R&U) If a social choice function satisfies the Pareto criterion, it must satisfy the unanimity criterion. Why? From class: Suppose that candidate X gets every firstplace vote. Can any other candidate Y win? Nope! If Y is any other candidate, then every voter places X above Y (since Y gets all first-place votes). Thus Y is not Pareto optimal, so Y cannot win if our method is Pareto. Since someone must win (social choice functions cannot be indifferent), it must be X. 10

The Pareto criterion (cont d) Proposition: The plurality method satisfies the Pareto criterion. Why? Is dictatorship Pareto? Monarchy? 11

The Pareto criterion (cont d) Proposition: The plurality method satisfies the Pareto criterion. Proof (from class): Suppose that we are given a profile in which every voter ranks candidate X over candidate Y. We must show that Y cannot win in the plurality method. How does the plurality method pick winners? By finding the candidate with the most first-place votes. But in this profile, candidate Y gets no first-place votes, because they are always ranked behind X! Thus, candidate Y cannot win in the plurality method. QED 12

The Pareto criterion (cont d) Dictatorships are Pareto: If candidate X is ranked above candidate Y in every profile, then in particular, this is true in the dictator s profile, so Y cannot be the dictator s first choice, and thus does not win. Monarchies are not Pareto: Consider the following counterexample: Suppose A is the monarch, but we are given the profile B B C C C B A A A A is the winner, but is clearly not Pareto optimal. 13

The Pareto criterion (cont d) Proposition: The anti-plurality method does not satisfy the Pareto criterion. Consider the following counterexample (from class): A A A B B B C C C In the anti-plurality method, A and B are both winners, but B is not Pareto optimal (since A is unanimously preferred). 14

The Condorcet candidate Given a profile, a candidate is called a Condorcet candidate if he or she beats every other candidate in a head-to-head (simple majority) match-up. Note, not every profile has a Condorcet candidate, for example: A C B B A C C B A Nicolas de Condorcet (1743-1794) Consequently, the method, select the Condorcet candidate as the winner, is not actually a social choice function. However, if there is a Condorcet candidate, they are unique. Why? If there were two Condorcet candidates, each would defeat the other in a head-to-head match-up, which is impossible. 15

Condorcet criteria A social choice function satisfies the Condorcet criterion if whenever a profile has a Condorcet candidate, the methods selects that candidate as the unique winner. Similar in spirit: A candidate is the anti-condorcet candidate if he or she loses to every other candidate in a head-to-head match-up. The same example shows that not every profile has an anti-condorcet candidate. A social choice function satisfies the anti-condorcet criterion if whenever a profile has an anti-condorcet candidate, the methods must not select that candidate as a winner. 16

Condorcet criteria (cont d) (Exercise 3.1 in R&U) Consider the profile A D E C D B B B A E C E A B C D A C E B E C D D A We claim that B is the Condorcet candidate. Why? Consider the head-to-head match-ups: 3 2 3 2 3 2 3 2 B A B C B D B E A B C B D B E B Since B wins each of these match-ups, B is the Condorcet candidate 17

Condorcet criteria (cont d) (Exercise 3.3 in R&U) Can you come up with a profile that has an anti-condorcet candidate, but no Condorcet candidate? Consider the example profile (from class): A B A B B A B A C C C C C is anti-condorcet (losing to A and B 4-0), but neither A nor B are Condorcet (since they tie 2-2). 18

Condorcet criteria (cont d) Proposition: The plurality method does not satisfy the Condorcet nor the anti-condorcet criteria. Counterexamples? Dictatorship? Monarchy? 19

Condorcet criteria (cont d) Proposition: The plurality method does not satisfy the Condorcet nor the anti-condorcet criteria. Proof (differs from class): Consider the following tabulated profile: 3 2 2 A B C B C B C A A A is the unique plurality winner. However, B is the Condorcet candidate since B beats A 4-3 and beats C 5-2. Thus, this method fails the Condorcet criterion. In fact, A is anti-condorcet in this example! She loses 3-4 to C as well. Thus, this example also shows that this method fails the anti-condorcet criterion. QED 20

Condorcet criteria (cont d) Dictatorships are not Condorcet or anti-condorcet: Use the profile on the previous slide, with A as the first voter as the dictator. Then A wins, despite B being the Condorcet candidate, and A being anti-condorcet. Monarchies are not Condorcet or anti-condorcet: Use the same example with A the monarch. 21

Condorcet criteria (cont d) Proposition: Copeland s method satisfies both the Condorcet and anti-condorcet criteria. Proof (from class): We ll do Condorcet first. Suppose we are given a profile with a Condorcet candidate X. We must show that X is the unique winner with Copeland s method. Since X beats every opponent in a head-to-head match-up, X gets the maximum possible number of points in Copeland s method (which is n-1, if there are n candidates, one for each match X wins). (R&U calls this a perfect score.) Meanwhile, every other candidate gets at least one fewer points, that is, no more than n-2 points, since they lose at least one match-up (namely, the one against X). Thus, X must be the unique winner. To see that Copeland satisfies anti-condorcet, note that an anti-condorcet candidate Y gets 0 points, since they lose every match-up, and the other candidates get at least 1 point (from the match against Y), so Y cannot win. QED 22

Recommended reading: Section 3.3 in R&U Problem set #2 is due tomorrow in class. Solutions to Problem set #1 have been posted on the course website. 23