Abstract
The main purpose of this paper is to compute the theoretical likelihood of some electoral outcomes under the impartial anonymous culture in four-candidate elections by using the last versions of software like LattE or Normaliz. By comparison with the three-candidate case, our results allow to analyze the impact of the number of candidates on the occurrence of these voting outcomes.
Similar content being viewed by others
Notes
Recall that we only consider large electorates; consequently, we ignore here the cases where two candidates obtain the same score: for instance, if the score of a is not strictly higher than the score of b, it means that the score of c is strictly lower than the score of c.
References
Barvinok, A. (1994). A polynomial time algorithm for counting integral points in polyhedra when the dimension is fixed. Mathematics of Operations Research,19, 769–779.
Brandt, F., Geist, C., & Strobel, M. (2016). Analyzing the practical relevance of voting paradoxes via Ehrhart theory, computer simulations and empirical data. In J. Thangarajah et al. (Eds.) Proceedings of the 15th international conference on autonomous agents and multiagent systems (AAMAS 2016).
Brandt, F., Hofbauer, J., & Strobel, M. (2019). Exploring the no-show paradox for Condorcet extensions using Ehrhart theory and computer simulations. http://dss.in.tum.de/files/brandt-research/noshow_study.pdf. Accessed 15 Oct 2019.
Bruns, W., & Ichim, B. (2018). Polytope volumes by descent in the face lattice and applications in social choice. arXiv preprint arXiv:1807.02835.
Bruns, W., Ichim, B., & Söger, C. (2019). Computations of volumes and Ehrhart series in four candidate elections. Annals of Operations Research, 280, 241–265.
Bruns, W., & Söger, C. (2015). The computation of generalized Ehrhart series in Normaliz. Journal of Symbolic Computation,68, 75–86.
Bubboloni, D., Diss, M., & Gori, M. (2018). Extensions of the Simpson voting rule to committee selection setting, forthcoming in Public Choice.
Clauss, P., & Loechner, V. (1998). Parametric analysis of polyhedral iteration spaces. Journal of VLSI Signal Processing, 19(2), 179–194.
De Loera, J. A., Dutra, B., Koeppe, M., Moreinis, S., Pinto, G., & Wu, J. (2013). Software for exact integration of polynomials over polytopes. Computational Geometry: Theory and Applications,46, 232–252.
De Loera, J. A., Hemmecke, R., Tauzer, J., & Yoshida, R. (2004). Effective lattice point counting in rational convex polytopes. Journal of Symbolic Computation,38, 1273–1302.
Diss, M., & Doghmi, A. (2016). Multi-winner scoring election methods: Condorcet consistency and paradoxes. Public Choice,169, 97–116.
Diss, M., Kamwa, E., & Tlidi, A. (2018). A note on the likelihood of the absolute majority paradoxes. Economic Bulletin,38, 1727–1734.
Diss, M., Kamwa, E., & Tlidi, A. (2019). On some k-scoring rules for committees elections: agreement and Condorcet principle. https://hal.univ-antilles.fr/hal-02147735/document. Accessed 15 Oct 2019.
Diss, M., & Mahajne, M. (2019). Social acceptability of Condorcet committees. https://halshs.archives-ouvertes.fr/halshs-02003292/document. Accessed 15 Oct 2019.
Ehrhart, E. (1962). Sur les polyèdres rationnels homothétiques à n dimensions. Comptes Rendus de l’académie des Sciences paris,254, 616–618.
Gehrlein, W. V. (2001). Condorcet winners on four candidates with anonymous voters. Economics Letters,71, 335–340.
Gehrlein, W. V. (2002). Obtaining representations for probabilities of voting outcomes with effectively unlimited precision integer arithmetic. Social Choice and Welfare,19, 503–512.
Gehrlein, W. V. (2006). Condorcet’s paradox. Berlin: Springer.
Gehrlein, W. V., & Fishburn, P. C. (1976). The probability of the paradox of voting: a computable solution. Journal of Economic Theory,13, 14–25.
Gehrlein, W. V., & Lepelley, D. (2011). Voting paradoxes and group coherence. Berlin: Springer.
Gehrlein, W. V., & Lepelley, D. (2017). Elections, voting rules and paradoxical outcomes. Berlin: Springer.
Gehrlein, W. V., Lepelley, D., & Plassmann, F. (2018). An evaluation of the benefit of using two-stage election procedures. Homo Oeconomicus,35, 53–79.
Huang, H. C., & Chua, V. (2000). Analytical representation of probabilities under the IAC condition. Social Choice and Welfare,17, 143–155.
Lepelley, D. (1989). Contribution à l’analyse des procédures de décision collective, unpublished dissertation, université de Caen.
Lepelley, D., Louichi, A., & Smaoui, H. (2008). On Ehrhart polynomials and probability calculations in voting theory. Social Choice and Welfare,30, 363–383.
Lepelley, D., & Mbih, B. (1987). The proportion of coalitionally unstable situations under the plurality rule. Economics Letters,24, 311–315.
Moulin, H. (1988). Axioms of cooperative decision making. Cambridge: Cambridge University Press.
Schürmann, A. (2013). Exploiting polyhedral symmetries in social choice. Social Choice and Welfare,40, 1097–1110.
Verdoolaege, S., & Bruynooghe, M. (2008). Algorithms for weighted counting over parametric polytopes: a survey and a practical comparison. In Proceedings of the 2008 international conference on information theory and statistical learning (ITSL).
Wilson, M. C., & Pritchard, G. (2007). Probability calculations under the IAC hypothesis. Mathematical Social Sciences,54, 244–256.
Software
Barvinok by Verdoolaege S, ver. 0.34. (2011). http://freshmeat.net/projects/barvinok.
LattE integrale by De Loera, J.A., Hemmecke, R., Tauzer, J., Yoshida, R., & Köppe M., ver. 1.7.3. (2016). http://www.math.ucdavis.edu/~latte/.
Normaliz by Bruns W, Ichim B, and Söger C, ver. 3.6.2. (2018). http://www.mathmatik.uni-osnabrueck.de/normaliz.
lrs by Avis D, ver. 6.2. (2016). cgm.cs.mcgill.ca/~avis/C/lrs.html.
Convex, a Maple package for convex geometry, Franz. (2017). http://www.math.uwo.ca/faculty/franz/convex/.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: Computation of \( \Pr \left( {E_{3} ,\infty } \right) \)
Appendix: Computation of \( \Pr \left( {E_{3} ,\infty } \right) \)
Let \( \left( {E_{3} , n} \right) \) be the set of all voting situations, of size \( n \), in which \( E_{3} \) occurs. Since \( a \) must be first or second in the first stage of the sincere vote, by symmetry, we can write:
where \( \left( {G,n} \right) \) is the set of the voting situations in \( \left( {E_{3} , n} \right) \) for which \( a \) is first and \( b \) is second in the first stage, and \( \left( {H,n} \right) \) the set of the voting situations in \( \left( {E_{3} , n} \right) \) for which \( a \) is second and \( b \) is first in the first stage. Considering all possibilities for the choice of the candidate who wins after manipulation and the candidate who goes with him to the second stage, we obtain:
Here, for \( \alpha \), \( \beta \) in \( \left\{ {b,c,d} \right\} \) and \( \alpha \ne \beta \), the notation \( \left( {G^{\alpha \beta } ,n} \right) \) (resp. \( \left( {H^{\alpha \beta } ,n} \right) \)) denotes the subset of \( \left( {G,n} \right) \) (resp. \( \left( {H,n} \right) \)) of voting situations where \( \alpha \) and \( \beta \) go to the second stage after manipulation (in favor of \( \alpha \)) and \( \alpha \) beats \( \beta \) by a majority of votes. For simplicity, in what follows, these subsets will be denoted by \( G^{\alpha \beta } \) (resp. \( H^{\alpha \beta } \)).
Using Proposition 2 and deleting the redundant inequalities, it follows that the voting situations \( x \) in \( G^{bc} \), \( G^{cb}, \) and \( G^{cd} \) are characterized by the following parametric linear systems:
By symmetry between candidates \( c \) and \( d \), the systems characterizing \( G^{bd} \), \( G^{db}, \) and \( G^{dc} \) are obtained by permuting \( c \) and \( d \) in \( S_{n}^{bc} , S_{n}^{cb} {\text{and }}S_{n}^{cd}, \) respectively; therefore, we have \( \left| {G^{bd} } \right| = \left| {G^{bc} } \right| \), \( \left| {G^{db} } \right| = \left| {G^{cb} } \right|, \) and \( \left| {G^{dc} } \right| = \left| {G^{cd} } \right| \).
Now, we use (12) and we apply the inclusion–exclusion principle to calculate \( \left| {\left( {G,n} \right)} \right| \). For the 15 pairwise intersections, it is obvious that \( G^{bc} \cap G^{cb} \), \( G^{bd} \cap G^{db}, \) and \( G^{cd} \cap G^{dc} \) are empty, and that by symmetry, we have \( \left| {G^{cb} \cap G^{bd} } \right| = \left| {G^{bc} \cap G^{db} } \right| \), \( \left| {G^{bd} \cap G^{dc} } \right| = \left| {G^{bc} \cap G^{cd} } \right| \), \( \left| {G^{bd} \cap G^{cd} } \right| = \left| {G^{bc} \cap G^{dc} } \right| \), \( \left| {G^{db} \cap G^{dc} } \right| = \left| {G^{cb} \cap G^{cd} } \right| \), and \( \left| {G^{db} \cap G^{cd} } \right| = \left| {G^{cb} \cap G^{dc} } \right| \). Of the 20 triple intersections, the only ones that are (possibly) non-empty are the 8 that are obtained by choosing one and only one element in each of the three sets \( \left\{ {G^{bc} , G^{cb} } \right\} \), \( \left\{ {G^{bd} , G^{db} } \right\} \) and \( \left\{ {G^{cd} , G^{dc} } \right\} \); and by symmetry we have \( \left| {G^{bc} \cap G^{bd} \cap G^{cd} } \right| = \left| {G^{bc} \cap G^{bd} \cap G^{dc} } \right| \), \( \left| {G^{bc} \cap G^{db} \cap G^{cd} } \right| = \left| {G^{cb} \cap G^{bd} \cap G^{dc} } \right| \), \( \left| {G^{cb} \cap G^{bd} \cap G^{cd} } \right| = \left| {G^{bc} \cap G^{db} \cap G^{dc} } \right|, \) and \( \left| {G^{cb} \cap G^{db} \cap G^{cd} } \right| = \left| {G^{cb} \cap G^{db} \cap G^{dc} } \right| \). Finally, all intersections of 4, 5, or 6 subsets \( G^{\alpha \beta } \) (\( \alpha \), \( \beta \) in \( \left\{ {b,c,d} \right\} \) and \( \alpha \ne \beta \)) are empty, because each of them is included in (at least) one of the three empty intersections, \( G^{bc} \cap G^{cb} \), \( G^{bd} \cap G^{db}, \) and \( G^{cd} \cap G^{dc} \) (to form an intersection of 4, 5 or 6 subsets \( G^{\alpha \beta } \), it is necessary to choose the two elements of at least one of the sets \( \left\{ {G^{bc} , G^{cb} } \right\} \), \( \left\{ {G^{bd} , G^{db} } \right\} \) and \( \left\{ {G^{cd} , G^{dc} } \right\} \)).
We can now write the formula giving the cardinality of \( \left( {G,n} \right){:} \)
To obtain \( { Pr }\left( {G,\infty } \right) \), we replace each cardinality that appears in the second member of (14) by the volume of the associated polytope (for example, the polytope associated with \( G^{bc} \) is the one described by the system \( S_{1}^{bc} \)), and then, we divide by the volume associated with the total number of voting situations (i.e., by\( 1/23! \)). Using the method based on LattE and Lrs (and Normaliz for the triple intersections), we get the following results:
Volume of the associated polytope | |
---|---|
\( G^{bc} \) | 21579965102214833661822341895472578264296147661301820748996857636891534517195601│1696627485451304216801706693769459989090225696942760631389397188608000000000000000 |
\( G^{cb} \) | 5573287256158164544822279100674154863906571658254932679572826363232135917│5381633372644158916218281473636071209286170854780673282015232000000000000000 |
\( G^{cd} \) | 690763588675926208892200201275635242424317217116771810711456400097168422586192409│1085841590688834698753092284012454393017744446043366804089214200709120000000000000000 |
\( G^{bc} \cap G^{bd} \) | 4374958709307366498551275282550829945168341340158005803124163906777355886621592418407│15783821593121799418221188750549477369596125747406423643972960256000000000000000000000000 |
\( G^{cb} \cap G^{db} \) | 371668307604755922004052912809417026311491652333420360381433590516107912320558408681833214907876615097308537557287│149865786232755160714603713492567144588895472965955187951797051990988320128219435626412335975849000960000000000000000 |
\( G^{bc} \cap G^{db} \) | 73728795097691910608032675802979649504177582112609999559155361106901059950083499782664269443828800493654344962886188347693│319769506827021125451226886147064465003853770355663104931071786212847273139053736326218431673677463018274816000000000000000000 |
\( G^{bc} \cap G^{cd} \) | 1128678697638588406020837480381418976300873421867855410386970684729753714450967582951899081133195953970779591091254147647624909│5730269562340218568085985799755395212869059564773482840364806408934223134651842954965834295592300137287484702720000000000000000000 |
\( G^{bc} \cap G^{dc} \) | 1791208843746302469037817356603077744950568899273425055696702056165779778059081619821156647740788159470371527351823357│7196555054896902817515270321913074283158760611825168125445294436607259132557097298780320373560269026099200000000000000000 |
\( G^{cb} \cap G^{cd} \) | 156090914075338229952814649806280933859900502452413621807273601730707932075993177383819│104173222514603876160259845753626550639334429932882396050221537689600000000000000000000000 |
\( G^{cb} \cap G^{dc} \) | 16987196522687599537363832144822196085155735807228425387516053556339294374740663419980423405997963504376939242487320875885987│116944276782453440165020118362355004344266521730071078374791967529269859890853937856445597869230615046683361280000000000000000000 |
\( G^{bc} \cap G^{bd} \cap G^{cd} \) | 674324033423332239489570724345486219235182962634052751321547924919297782590615439333218137279830648518311866993975500318502409742149903│800337852832421947317129795594046017972207282304842286498696861526387919674178698272333760160140951092350184062976000000000000000000000000 |
\( G^{bc} \cap G^{db} \cap G^{dc} \) | 66224084947732991948095057625409755209795263748146831963545633222617962161248153298557211890052296338881958196326711071188575733861237│1026074170297976855534781789224877694611821446449338754679320110452101015342843422855427405148736019370814126161920000000000000000000000000 |
\( G^{cb} \cap G^{db} \cap G^{dc} \) | 172499040886591309105147664280974842140465399609634373368829431585429271796091241844107326681316968683271609632632217239661825101449│3374674704134010572259781563482056846842725283481549285924564370689149907098236927927278529330469282801631887360000000000000000000000000 |
\( G^{cb} \cap G^{bd} \cap G^{dc} \) | 6433925339577046922056349489616669597071348937361564870673006196214723723│147708854420427693761988302384320276840921234854209315143680000000000000000000 |
After calculation, we obtain:
To compute \( { Pr }\left( {H, \infty } \right) \), we use (13) and we proceed in exactly the same way as for \( { Pr }\left( {G,\infty } \right) \). Using Proposition 2 and deleting the redundant inequalities, we obtain the systems characterizing the voting situations \( x \) in \( H^{bc} \), \( H^{cb}, \) and \( H^{cd} \):
Here again, \( c \) and \( d \) being symmetrical, we obtain the same symmetries as before. Therefore, the formula describing the cardinality of \( \left( {H,n} \right) \) is exactly the same as (14), except that \( \left| {\left( {G,n} \right)} \right| \) is replaced with \( \left| {\left( {H,n} \right)} \right| \) and each \( G^{\alpha \beta } \) (\( \alpha \), \( \beta \) in \( \left\{ {b,c,d} \right\} \) and \( \alpha \ne \beta \)) is replaced with \( H^{\alpha \beta } \). By applying the method based on LattE and Lrs (and Normaliz for the triple intersections), we calculate the volumes of all the polytopes associated with the cardinalities involved in the second member of this formula. We then get:
Volume of the associated polytope | |
---|---|
\( H^{bc} \) | 1122570228285484416840414329038859444597│22032517669125135387328512000000000000000 |
\( H^{cb} \) | 4815102613831086008681845331774767357│2754064708640641923416064000000000000000 |
\( H^{cd} \) | 4859814977353382934768278656966765458531281829079037659495505447570096549131│2636152533698047281461986677988445611214432454458846070397992960000000000000000 |
\( H^{bc} \cap H^{bd} \) | 6833482604299574922319228350257359109218343485126709565221│228793803992037731981659487629131907072000000000000000000000 |
\( H^{cb} \cap H^{db} \) | 5049881685008345434475345496190114508046016393│10408519742515465059176289612595200000000000000000 |
\( H^{bc} \cap H^{db} \) | 675827524940449071804022414564519275245666212472817163│831977469061955389811512540956047966208000000000000000000 |
\( H^{bc} \cap H^{cd} \) | 54527549003500598877896693999359561653981068895014160097048506522520901040183861295097633826231569261│4610787019086802067117459185013926303116690035637510458825406244812151254034364026060800000000000000000 |
\( H^{bc} \cap H^{dc} \) | 8308091188337554623944084932295328721407928052921637390327492939829616693208962854674889│6253787586569434428002409935673694836855808351567447350530718940594490572800000000000000000 |
\( H^{cb} \cap H^{cd} \) | 536467444833359046580624594631977457827077563033371231007512557540128954421668859753948816150778793│1129430486744758491847310206009682124024272495502035679704440095241071735752097792000000000000000000000 |
\( H^{cb} \cap H^{dc} \) | 8829802817815659461561114981596342439514389717111976228094675237975903263366333445667│24510239414342286607887164160978619779956136984391328044408069530058752000000000000000000 |
\( H^{bc} \cap H^{bd} \cap H^{cd} \) | 6526360468678669861037589052034470126857779241331314839826230041617566152568440550895761885023785390756518058117863413│7216599736503166408006324991787941691613794336591761487594800388300067534635371463393316550373015552000000000000000000000 |
\( H^{bc} \cap H^{db} \cap H^{dc} \) | 987760979695491729713131336886222849186173537468343019201090980213224186369431218746360706439232516885140014895617│3299771255831351809787985821576562273257336230723256281479103972702362841625684253952133767888896000000000000000000000 |
\( H^{cb} \cap H^{db} \cap H^{dc} \) | 5404398077528054472521504460835385332865736755704757514533585222718698855027095804933932537271634529787736605385657│38183067388905642370403835935385934876263462098369108399972488826984484310240060652874690742714368000000000000000000000 |
\( H^{cb} \cap H^{bd} \cap H^{dc} \) | 141767031443658485490372638618892830741501772123401878931601845278076851988166862281062607360239016458736050692837│763661347778112847408076718707718697525269241967382167999449776539689686204801213057493814854287360000000000000000000 |
After calculation, we obtain:
Finally, going to the limit in (11), we have:
Rights and permissions
About this article
Cite this article
El Ouafdi, A., Lepelley, D. & Smaoui, H. Probabilities of electoral outcomes: from three-candidate to four-candidate elections. Theory Decis 88, 205–229 (2020). https://doi.org/10.1007/s11238-019-09724-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11238-019-09724-5