Elections and Strategic Voting: Condorcet and Borda E. Maskin Harvard University Lecture II Gorman Lectures University College, London February, 2016 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) 2 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) • prominent examples 3 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) • prominent examples – Plurality Rule (MPs in Britain, members of Congress in U.S.) 4 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) • prominent examples – Plurality Rule (MPs in Britain, members of Congress in U.S.) choose candidate ranked first by more voters than any other 5 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) • prominent examples – Plurality Rule (MPs in Britain, members of Congress in U.S.) choose candidate ranked first by more voters than any other – Majority Rule (Condorcet Method) 6 • Today (as yesterday) will discuss voting rules methods for choosing winning candidate on basis of voters’ preferences (rankings, utility functions) • prominent examples – Plurality Rule (MPs in Britain, members of Congress in U.S.) choose candidate ranked first by more voters than any other – Majority Rule (Condorcet Method) choose candidate preferred by majority to each other candidate 7 − Run-off Voting (presidential elections in France) 8 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority 9 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority 10 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority − Rank-Order Voting (Borda Count) 11 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority − Rank-Order Voting (Borda Count) • candidate assigned 1 point every time some voter ranks her first, 2 points every time ranked second, etc. 12 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority − Rank-Order Voting (Borda Count) • candidate assigned 1 point every time some voter ranks her first, 2 points every time ranked second, etc. • choose candidate with lowest point total 13 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority − Rank-Order Voting (Borda Count) • candidate assigned 1 point every time some voter ranks her first, 2 points every time ranked second, etc. • choose candidate with lowest point total − Utilitarian Principle 14 − Run-off Voting (presidential elections in France) • choose candidate ranked first by more voters than any other, unless number of first-place rankings less than majority among top 2 candidates, choose alternative preferred by majority − Rank-Order Voting (Borda Count) • candidate assigned 1 point every time some voter ranks her first, 2 points every time ranked second, etc. • choose candidate with lowest point total − Utilitarian Principle • choose candidate who maximizes sum of voters’ utilities 15 • Which voting rule to adopt? 16 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule 17 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy 18 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them 19 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them • One important criterion: nonmanipulability 20 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them • One important criterion: nonmanipulability – voters shouldn’t have incentive to misrepresent preferences, i.e., vote strategically 21 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them • One important criterion: nonmanipulability – voters shouldn’t have incentive to misrepresent preferences, i.e., vote strategically – otherwise 22 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them • One important criterion: nonmanipulability – voters shouldn’t have incentive to misrepresent preferences, i.e., vote strategically – otherwise not implementing intended voting rule 23 • Which voting rule to adopt? • As discussed yesterday, answer depends on what one wants in voting rule – can specify criteria (axioms) voting rule should satisfy – see which rules best satisfy them • One important criterion: nonmanipulability – voters shouldn’t have incentive to misrepresent preferences, i.e., vote strategically – otherwise not implementing intended voting rule decision problem for voters may be hard 24 • But basic negative result 25 • But basic negative result Gibbard-Satterthwaite (GS) theorem 26 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) 27 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic 28 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic – requires that voting rule never be manipulable 29 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic – requires that voting rule never be manipulable – but some circumstances where manipulation can occur may be unlikely 30 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic – requires that voting rule never be manipulable – but some circumstances where manipulation can occur may be unlikely • In any case, natural question: 31 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic – requires that voting rule never be manipulable – but some circumstances where manipulation can occur may be unlikely • In any case, natural question: Which (reasonable) voting rule(s) nonmanipulable most often? 32 • But basic negative result Gibbard-Satterthwaite (GS) theorem – if 3 or more candidates, no voting rule is always nonmanipulable (except for dictatorial rules - - where one voter has all the power) • Still, GS overly pessimistic – requires that voting rule never be manipulable – but some circumstances where manipulation can occur may be unlikely • In any case, natural question: Which (reasonable) voting rule(s) nonmanipulable most often? • Lecture tries to answer question 33 • X = finite set of candidates 34 • X = finite set of candidates • society consists of a continuum of voters [0,1] 35 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 36 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon 37 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon • utility function for voter i U i : X 38 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon • utility function for voter i U i : X – restrict attention to strict utility functions 39 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon • utility function for voter i U i : X – restrict attention to strict utility functions if x y, then U i x U i y 40 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon • utility function for voter i U i : X – restrict attention to strict utility functions if x y, then U i x U i y U X = set of strict utility functions 41 • X = finite set of candidates • society consists of a continuum of voters [0,1] – typical voter i 0,1 – reason for continuum clear soon • utility function for voter i U i : X – restrict attention to strict utility functions if x y, then U i x U i y U X = set of strict utility functions • profile U - - specification of each individual's utility function 42 • voting rule F 43 • voting rule F for all profiles U and all Y X , F U , Y Y 44 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot 45 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U 46 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) 47 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate 48 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled 49 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties 50 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie 51 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most 52 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most – with rank-order voting, might be two candidates who each get lowest number of points 53 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most – with rank-order voting, might be two candidates who each get lowest number of points • But exact ties unlikely with many voters 54 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most – with rank-order voting, might be two candidates who each get lowest number of points • But exact ties unlikely with many voters – with continuum, ties are nongeneric 55 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most – with rank-order voting, might be two candidates who each get lowest number of points • But exact ties unlikely with many voters – with continuum, ties are nongeneric • so, correct definition of decisiveness: 56 • voting rule F for all profiles U and all Y X , F U , Y Y Y is ballot F U , Y optimal candidate(s) in Y if profile is U • will require F to be decisive (D) F U , Y single candidate – reflects fact that there is single office to be filled • definition isn’t quite right - - ignores ties – with any voting rule, there could be exact tie – with plurality rule, might be two candidates who are both ranked first the most – with rank-order voting, might be two candidates who each get lowest number of points • But exact ties unlikely with many voters – with continuum, ties are nongeneric • so, correct definition of decisiveness: for generic profile U and all Y X F U , Y is singleton 57 plurality rule: 58 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a 59 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a majority rule: 60 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a majority rule: F C U , Y a i U i a U i b 1 2 for all b 61 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a majority rule: F C U , Y a i U i a U i b 1 2 for all b rank-order voting: 62 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a majority rule: F C U , Y a i U i a U i b 1 2 for all b rank-order voting: F B U , Y a rUi a d i rUi b d i for all b , where rUi a # b U i b U i a 63 plurality rule: F P U , Y a i U i a U i b for all b i U i a U i b for all b for all a majority rule: F C U , Y a i U i a U i b 1 2 for all b rank-order voting: F B U , Y a rUi a d i rUi b d i for all b , where rUi a # b U i b U i a utilitarian principle: F U U , Y a U i a d i U i b d i for all b 64 What properties should reasonable voting rule satisfy? 65 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y 66 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y – if everybody prefers x to y, y should not be chosen 67 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y – if everybody prefers x to y, y should not be chosen • Anonymity (A): suppose : 0,1 0,1 measure-preserving permutation. If Ui U (i ) for all i, then 68 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y – if everybody prefers x to y, y should not be chosen • Anonymity (A): suppose : 0,1 0,1 measure-preserving permutation. If Ui U (i ) for all i, then F U , Y F U , Y for all Y 69 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y – if everybody prefers x to y, y should not be chosen • Anonymity (A): suppose : 0,1 0,1 measure-preserving permutation. If Ui U (i ) for all i, then F U , Y F U , Y for all Y – candidate chosen depends only on voters’ preferences and not who has those preferences 70 What properties should reasonable voting rule satisfy? • Consensus / Pareto Property (P): if Ui x U i y for all i and x Y , then y F U , Y – if everybody prefers x to y, y should not be chosen • Anonymity (A): suppose : 0,1 0,1 measure-preserving permutation. If Ui U (i ) for all i, then F U , Y F U , Y for all Y – candidate chosen depends only on voters’ preferences and not who has those preferences – voters treated symmetrically 71 • Neutrality (N): Suppose : Y Y permutation. 72 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, 73 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then 74 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . 75 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically 76 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically • All four voting rules – plurality, majority, rank-order, utilitarian – satisfy P, A, N 77 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically • All four voting rules – plurality, majority, rank-order, utilitarian – satisfy P, A, N • Next axiom most controversial 78 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically • All four voting rules – plurality, majority, rank-order, utilitarian – satisfy P, A, N • Next axiom most controversial still 79 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically • All four voting rules – plurality, majority, rank-order, utilitarian – satisfy P, A, N • Next axiom most controversial still • has quite compelling justification 80 • Neutrality (N): Suppose : Y Y permutation. If U i ,Y x >U i ,Y y U i x U i y for all x, y, i, then F U ,Y , Y F U , Y . – candidates treated symmetrically • All four voting rules – plurality, majority, rank-order, utilitarian – satisfy P, A, N • Next axiom most controversial still • has quite compelling justification • invoked by both Arrow (1951) and Nash (1950) 81 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y 82 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y then 83 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y then x F U , Y 84 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y then x F U , Y – if x chosen and some non-chosen candidates removed, x still chosen 85 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y then x F U , Y – if x chosen and some non-chosen candidates removed, x still chosen – Nash formulation (rather than Arrow) 86 • No Spoilers / Independence of Irrelevant Candidates (I): if x F U , Y and x Y Y then x F U , Y – if x chosen and some non-chosen candidates removed, x still chosen – Nash formulation (rather than Arrow) – no “spoilers” (e.g. Nader in 2000 U.S. presidential election, Le Pen in 2002 French presidential election) 87 • Majority rule and utilitarianism satisfy I, but others don’t: 88 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule 89 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x 90 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x 91 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x F P U ,x, y y 92 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x F P U ,x, y y – rank-order voting 93 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x F P U ,x, y y – rank-order voting .55 x y z .45 y z x 94 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x F P U ,x, y y – rank-order voting .55 x y z .45 y z x F B U ,x, y, z y 95 • Majority rule and utilitarianism satisfy I, but others don’t: – plurality rule .35 x y z .33 y z x .32 z y x F P U ,x, y, z x F P U ,x, y y – rank-order voting .55 x y z .45 y z x F B U ,x, y, z y F B U ,x, y x 96 Final Axiom: 97 Final Axiom: • Nonmanipulability (NM): 98 Final Axiom: • Nonmanipulability (NM): if x F U , Y and x F U , Y , where U j U j for all j C 0,1 99 Final Axiom: • Nonmanipulability (NM): if x F U , Y and x F U , Y , where U j U j for all j C 0,1 then 100 Final Axiom: • Nonmanipulability (NM): if x F U , Y and x F U , Y , where U j U j for all j C 0,1 then U i x U i x for some i C 101 Final Axiom: • Nonmanipulability (NM): if x F U , Y and x F U , Y , where U j U j for all j C 0,1 then U i x U i x for some i C – the members of coalition C can’t all gain from misrepresenting utility functions as U i 102 • NM implies voting rule must be ordinal (no cardinal information used) 103 • NM implies voting rule must be ordinal (no cardinal information used) • F is ordinal if whenever, for profiles U and U , U i x U i y U i x U i y for all i, x, y 104 • NM implies voting rule must be ordinal (no cardinal information used) • F is ordinal if whenever, for profiles U and U , U i x U i y U i x U i y for all i, x, y (*) F U , Y F U , Y for all Y 105 • NM implies voting rule must be ordinal (no cardinal information used) • F is ordinal if whenever, for profiles U and U , U i x U i y U i x U i y for all i, x, y (*) F U , Y F U , Y for all Y • Lemma: If F satisfies NM, F ordinal 106 • NM implies voting rule must be ordinal (no cardinal information used) • F is ordinal if whenever, for profiles U and U , U i x U i y U i x U i y for all i, x, y (*) F U , Y F U , Y for all Y • Lemma: If F satisfies NM, F ordinal • NM rules out utilitarianism 107 But majority rule also violates NM 108 But majority rule also violates NM • F C violates decisiveness 109 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y 110 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z 111 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle 112 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner 113 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility 114 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise 115 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise (Black's method) 116 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise – (Black's method) extensions make F C vulnerable to manipulation 117 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise – – (Black's method) extensions make F C vulnerable to manipulation .35 x y z .33 y z x .32 z x y 118 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise – – (Black's method) extensions make F C vulnerable to manipulation .35 x y z .33 y z x .32 z x y F C / B U , x, y, z x 119 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise – – (Black's method) extensions make F C vulnerable to manipulation .35 x y z .33 y z x .32 z x y F C / B U , x, y, z x z y x 120 But majority rule also violates NM • F C violates decisiveness .35 x y z .33 y z x .32 z x y F C U , x, y, z – example of Condorcet cycle – F C must be extended when no majority winner – one possibility F C U , Y , if nonempty F C / B U , Y F B U , Y , otherwise – – (Black's method) extensions make F C vulnerable to manipulation .35 x y z .33 y z x z y x .32 z x y F C / B U , x, y, z x F C / B U , x, y, z z 121 Theorem: There exists no voting rule satisfying D,P,A,N,I, and NM 122 Theorem: There exists no voting rule satisfying D,P,A,N,I, and NM Proof: similar to that of GS 123 Theorem: There exists no voting rule satisfying D,P,A,N,I, and NM Proof: similar to that of GS overly pessimistic - - many cases in which some rankings unlikely 124 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles 125 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles 126 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 127 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush 128 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush unlikely that many had ranking Bush Nader Gore Nader or Bush Gore 129 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush unlikely that many had ranking Bush Nader Gore • if you have ranking order Nader or Bush Gore Nader Gore Bush 130 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush unlikely that many had ranking Bush Nader Gore • if you have ranking order – would it pay to say Gore ? Nader or Bush Gore Nader Gore Bush Nader Bush 131 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush unlikely that many had ranking Bush Nader Gore • if you have ranking order – would it pay to say Gore ? Nader or Bush Gore Nader Gore Bush Nader Bush – no - just makes it more likely that Gore beats Nader 132 Lemma: Majority rule satisfies all 6 properties if and only if preferences restricted to domain with no Condorcet cycles When can we rule out Condorcet cycles? • preferences single-peaked 2000 US election Nader Gore Bush unlikely that many had ranking Bush Nader Gore • if you have ranking order – would it pay to say Gore ? Nader or Bush Gore Nader Gore Bush Nader Bush – no - just makes it more likely that Gore beats Nader – but prefer Nader 133 • strongly-felt candidate 134 • strongly-felt candidate – in 2002 French election, 3 main candidates: Chirac, Jospin, Le Pen 135 • strongly-felt candidate – in 2002 French election, 3 main candidates: Chirac, Jospin, Le Pen – voters didn’t feel strongly about Chirac and Jospin 136 • strongly-felt candidate – in 2002 French election, 3 main candidates: Chirac, Jospin, Le Pen – voters didn’t feel strongly about Chirac and Jospin – felt strongly about Le Pen (ranked him first or last) 137 Voting rule F works very well on domain U if satisfies D,P,A,N,I,NM when utility functions restricted to U 138 Voting rule F works very well on domain U if satisfies D,P,A,N,I,NM when utility functions restricted to U e.g., F C works very well when preferences single-peaked 139 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. 140 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . 141 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . Then if there exisits profile U on U C such that F U , Y F C U , Y for some Y , 142 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . Then if there exisits profile U on U C such that F U , Y F C U , Y for some Y , there exists domain U on which F C works well but F does not 143 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . Then if there exisits profile U on U C such that F U , Y F C U , Y for some Y , there exists domain U on which F C works well but F does not Proof: From NM, if F works very well on U , F must be ordinal 144 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . Then if there exisits profile U on U C such that F U , Y F C U , Y for some Y , there exists domain U on which F C works well but F does not Proof: From NM, if F works very well on U , F must be ordinal • Hence first half follows from Dasgupta-Maskin (2008) 145 • Theorem 1: Suppose F works very well on domainU , then F C works well on U too. C C • Conversely, suppose that F works well on U . Then if there exisits profile U on U C such that F U , Y F C U , Y for some Y , there exists domain U on which F C works well but F does not Proof: From NM, if F works very well on U , F must be ordinal • Hence first half follows from Dasgupta-Maskin (2008) – shows that Theorem 1 holds when NM replaced by ordinality and “works very well” replaced by “works well” 146 • Suppose F works very well on U 147 • Suppose F works very well on U • If F C doesn't work very well on U , Lemma impliesU must contain Condorcet cycle x y z y z x z x y 148 • Consider 149 • Consider 1 2 U1 x z z x n z x 150 • Consider 1 2 U1 x z z x (*) n z x Suppose F U 1 , x, z z 151 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y 2 n z x y 152 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y 2 n z x y F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 153 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y 2 n z x y F U , x, y, z y (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 2 154 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y 2 n z x y F U , x, y, z y (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 155 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y 2 n z x y F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 156 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • so F U 2 , y, z z (I) 157 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • • so F U 2 , y, z z (I) so for 158 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • • so F U 2 , y, z z (I) so for 1 2 3 U x x z z z x 3 n z x 159 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • • so F U 2 , y, z z (I) so for 1 2 3 U x x z z z x 3 n z x F U 3 , x, z z (N) 160 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • • so F U 2 , y, z z (I) so for 1 2 3 U x x z z z x 3 n z x F U 3 , x, z z (N) • 1 Continuing in the same way, let U x z 4 n 1 x z n z x 161 • Consider 1 2 U1 x z z x (*) • n z x Suppose F U 1 , x, z z 1 2 3 U x y z y z x z x y n z x y 2 F U , x, y, z y F U , x, y, z z (from I) F U , x, y y, contradicts (*) (A,N) F U 2 , x, y, z x (from I) F U 2 , x, z x, contradicts (*) 2 so 2 2 • • so F U 2 , y, z z (I) so for 1 2 3 U x x z z z x 3 n z x F U 3 , x, z z (N) • 1 Continuing in the same way, let U x z 4 n 1 x z n z x F U 4 , x, z z , contradicts (*) 162 • So F can’t work very well on U with Condorcet cycle 163 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y 164 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y • Then there exist with 1 and 165 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y • Then there exist with 1 and U 1 x y y x 166 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y • Then there exist with 1 and U 1 x y y x such that 167 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y • Then there exist with 1 and U 1 x y y x such that x F C U , x, y and y F U , x, y 168 • So F can’t work very well on U with Condorcet cycle C • Conversely, suppose that F works well on U C and F U , Y F C U , Y for some U and Y • Then there exist with 1 and U 1 x y y x such that x F C U , x, y and y F U , x, y • But not hard to show that F C unique voting rule satisfying D,P,A,N, and NM when X 2 - - contradiction 169 • Let’s drop I 170 • Let’s drop I – most controversial 171 • Let’s drop I – most controversial • no voting rule satisfies D,P,A,N,NM on U X 172 • Let’s drop I – most controversial • no voting rule satisfies D,P,A,N,NM on U X – GS again 173 • Let’s drop I – most controversial • no voting rule satisfies D,P,A,N,NM on U X – GS again • F works nicely on U if satisfies D,P,A,N,NM on U 174 Theorem 2: 175 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. 176 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . 177 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that 178 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , 179 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not 180 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: 181 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain 182 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain • F B works nicely only when U is subset of Condorcet cycle 183 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain • F B works nicely only when U is subset of Condorcet cycle C B • so F and F complement each other 184 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain • F B works nicely only when U is subset of Condorcet cycle C B • so F and F complement each other – if F works nicely on U and U doesn't contain Condorcet cycle, F C works nicely too 185 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain • F B works nicely only when U is subset of Condorcet cycle C B • so F and F complement each other – if F works nicely on U and U doesn't contain Condorcet cycle, F C works nicely too – if F works nicely on U and U contains Condorcet cycle, then U can't contain any other ranking (otherwise no voting rule works nicely) 186 Theorem 2: • Suppose F works nicely on U , then F C or F B works nicely on U too. Conversely, suppose F works nicely onU , where F F C or F B . Then, if there exisits profile U on U such that F U , Y F U , Y for some Y , there exists domain U on which F * works nicely but F does not Proof: • F C works nicely on any Condorcet-cycle-free domain • F B works nicely only when U is subset of Condorcet cycle C B • so F and F complement each other – if F works nicely on U and U doesn't contain Condorcet cycle, F C works nicely too – if F works nicely on U and U contains Condorcet cycle, then U can't contain any other ranking (otherwise no voting rule works nicely) – so F B works nicely on U . 187 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also 188 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also – only two that work nicely on maximal domains 189 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also – only two that work nicely on maximal domains • sense in which Condorcet is better than Borda 190 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also – only two that work nicely on maximal domains • sense in which Condorcet is better than Borda – if there are 3 alternatives, then 6 strict orderings, including 2 Condorcet cycles x y z x z y y z z x x y z y x y x z 191 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also – only two that work nicely on maximal domains • sense in which Condorcet is better than Borda – if there are 3 alternatives, then 6 strict orderings, including 2 Condorcet cycles x y z x z y y z x z y x z x y y x z – Borda works nicely on either cycle (domain of 3 orderings) 192 • Striking that the 2 longest-studied voting rules (Condorcet and Borda) are also – only two that work nicely on maximal domains • sense in which Condorcet is better than Borda – if there are 3 alternatives, then 6 strict orderings, including 2 Condorcet cycles x y z x z y y z x z y x z x y y x z – Borda works nicely on either cycle (domain of 3 orderings) – Condorcet works nicely as long as one ordering from each cycle missing (domain of 4 orderings) 193