Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Report on Rosie Hackett decision-final draft


Published on

  • Login to see the comments

  • Be the first to like this

Report on Rosie Hackett decision-final draft

  1. 1. DublinCity Council’sRosie Hackett Bridge:A Landmark in Decision-Making JohnBaker1 On 2 September, 2013, DublinCityCouncil votedtoname the newestbridge overthe RiverLiffeythe Rosie HackettBridge.Whatmakesthisa landmarkdecisionisthatitseemstohave beenthe first authoritative decisiontakenbyapublicbodyinIreland – and perhapseveninEurope – to have used the votingprocedure knownasthe Borda Count,referredtointhe Council’sproceedingsasa Preferendum (DublinCityCouncil2013a, item24). Thisreport summarisesthe process, analysesthe results,anddiscussessome of the technical issuesthatarise withthismethodof voting.Itconcludes that the procedure waswell suitedtothe taskinhand. Background The processfor namingthe bridge wasreferredtothe CommemorativeNamingCommittee chaired by CouncillorDermotLacey.AccordingtoLacey (personal communication),the committee agreed fromthe start that itwoulduse an open,participatoryprocesstofindaname,andinvited submissionsfromthe public.Inthe course of the process,itreceived thousandsof itemsof correspondence,andofficial applicationsfor85 names.This initial listwasnarroweddowntoabout thirtythrougha consensual process withinthe committee, startingby eliminatingnamesof people whowere still alive orhaddiedlessthantwentyyearsearlier,aswell asfigureswhohadalready beenhonoured byapublicnaming.The resultinglistwasthenfurtherreducedbydiscussionwithin the committee,leadinginstagestoa listof seventeenandthen ten(DublinCityCouncil 2013b). The listof tenwas reducedwithinthe committee toashortlistof five,usingaversionof aBorda Countvote among the six membersinattendance.The finalshortlistwasputto the full council, where all fifty-one membersparticipatedinaBorda Countvote. Detailsof these votes are given below. The Borda Count The Borda Count (hereafterBC) isa methodof votingnamedafterthe eighteenth-century French mathematician whodevelopedit,Jean-Charlesde Borda.Itis designedtopickawinner (orsetof winners) fromagroupof more than twooptionsor candidates.The basicBCmethodoperatesas follows:Eachvoterindicatestheir opiniononthe options by numberingthem1,2, 3… inorder of preference.If there are n options, avoter’sfirstpreference isgiven n points,theirsecondpreference n-1,and so on.All of the pointsare addedup andthe optionwiththe highesttotal isthe winner. The main advantage of the BC is thatit takesintoaccount voters’preferences amongthe complete setof options,andtherefore favourspoliciesthathave wide support.Bycontrast,majorityrule only considersfirstpreferences,and canthereby leadtodecisionsthatare stronglyopposedbyalarge minorityof the electorate. 1 Emeritus Professor,School of Social Justice,University CollegeDublin.I am grateful to Councillor Dermot Lacey and to Paula Ebbs in the Dublin City Council Secretariatfor their assistance,and to Peter Emerson and Phil Kearney for their comments and suggestions.
  2. 2. There are several variantsof the basicBC method.Forinstance, the best-known example in Europe of BC votingisthe EurovisionSongContest,whereeachcountryvotes for their top ten alternatives, and the top two of these are given 12 and 10 points instead of 10 and 9. One of the mostcommon questionsabout the BChas to do withincompleteballots.If, inafive- optioncontext,avoteronlyindicatestheirfirstpreference andthisvote isgiven5points, itgainsa 5-pointleadoverall the otheralternatives,whichseemsunfair.Inthe ModifiedBCsystem (MBC), developedbyEmerson (Emerson2012),the solutionistogive a voter’stoppreference mpoints, where mis the numberof options actually votedfor. Soif a voterhas indicated only theirfirsttwo preferences,these are given2pointsand1 point. The discussionbelow displays bothBCandMBC results. Use of the Borda Count by the Commemorative Naming Committee The committee usedaformof BC to reduce the shortlistfromtento five. Eachmemberindicated theirtopfive preferences onasecretballot.All six votersusedall fiveof theirpreferences.Inthe count,5 pointswasgiventoa firstpreference,4to a second,etc. The five optionswiththe highest scoreswere includedinthe final shortlist (DublinCityCouncil 2013b).Table 1 showsthe complete count andits result. Table 1. Committee vote. Numbers above the line are preferences. Selected shortlist in red. In Ireland,the question naturally arises of how the BC compares with Proportional Representation by Single Transferable Vote (PRSTV). This ballot shows a very clear difference. Since exactly five of the options received someone’s first preference, those five would have been chosen by PRSTV.2 Theyinclude AbbeyandSigerson insteadof Duff andStoker.Proponents of the BC argue that such a result neglects the clear evidence that Duff is much more popular with the electorate as a whole than eitherAbbeyorSigerson.Sincethe pointof thisdecisionis to reflect public opinion as a whole rather than to select public representatives, it makes more sense to use a procedure that reflects overall popularity than one, like PRSTV, that privileges first preferences. (The first-past-the-post result in this case would have been the same as PRSTV, with the same disadvantage.) The use of the Borda Count in full Council 2 The standard full method here would startby treating each vote as 100,setting the quota at101, and identifyingHackett’s surplus as 99.Sincethis is less than the 100 votes polled by four other options,those options would be deemed selected. voter Abbey Bermingh Connolly Duff Hackett Mills Sigerson Stoker Walton Yeats 1 3 5 1 4 2 2 5 1 3 4 2 3 2 4 3 1 5 4 2 4 5 1 3 5 1 3 2 5 4 6 2 3 1 4 5 BC score 6 19 4 15 14 10 7 8 3 4 rank 7 1 8= 2 3 4 6 5 10 8= 1st prefs 1 1 0 0 2 1 1 0 0 0
  3. 3. In the full Council meeting,all 51councillorswere invitedtoindicate theirpreferencesamongthe five shortlistedalternatives.Atthe meeting,the chair(LordMayor OisinQuinn) wasaskedif itwas necessaryto rank all five alternatives.He repliedthat voterswere notrequiredtodoso,but that theywouldbe answerable totheirconstituentsforwhateverwaytheyhadcompletedtheirballots, as these wouldbe made public. Onlytwelvecouncillorsreturnedincomplete ballots. ‘Rosie Hackett’ wonthe vote with192 points.Table 3 givesthe result of the ballot(‘BCscore’) andalsoshowswhat the resultwouldhave beenusingthe ModifiedBordaCount(‘MBCscore’). Inthiscase,the use of MBC wouldnothave affectedthe winner,butitwouldhave ledtoa tie forsecondplace. Table 3. Result of Council vote. For ‘BC score’, first preferences were given 5 points, second preferences 4 points, etc. For ‘MBC score’, see explanation in text. Because ‘Hackett’ had an overall majority of first preferences, it would have won the vote straightaway if it had been conducted by first-past-the-post, PRSTV or multiple-round voting.3 Analysis by party Because the ballotswere public,itispossible toobservedifferences invotingpatterns between and within political parties. Table 4 shows the basic BC result for the eight categories of councillors. Table 4. Breakdown by party / group, using basic BC. Top choice shown in red. The onlyparty groupthat unanimouslyrankedone optionfirst(namely‘Hackett’)wasSinnFéin; eveninthisgroup, secondandthirdpreferenceswere splitbetween‘Bermingham’and‘Mills’.The Labour Partywas nearlyunanimousinitsfirstpreferencesfor‘Hackett’(15 of 17 voters);itsvote became more fragmentedatlowerpreferences.The majorityof Fine Gael’sfirstpreferenceswentto 3 In multiple-round voting, voters choose one option in each round; the option with the lowest total is eliminated and another round takes place. Voting continues until one option has a majority of votes. This procedure is logically equivalentto PRSTV, provided voters do not change their preferences from one round to the next. I mention this procedure sinceitmight have been employed if the Council had not used BC (Lacey, personal communication). Bermingh Duff Hackett Mills Stoker first preferences 15 1 27 6 2 second preferences 11 1 7 25 6 third preferences 11 8 4 13 7 fourth preferences 7 16 5 3 11 fifth preferences 1 15 7 1 15 BC score 167 80 192 176 92 MBC score 156 76 165 156 90 Voters Bermingh Duff Hackett Mills Stoker Eirígí 1 3 1 5 4 2 Fianna Fáil 6 24 11 14 28 7 Fine Gael 12 51 24 24 41 39 Independent 8 31 14 34 25 15 Labour 17 37 25 80 60 23 People Before Profit 1 4 0 5 0 0 Sinn Féin 5 17 5 25 18 6 United Left 1 0 0 5 0 0
  4. 4. ‘Bermingham’andthe majorityof FiannaFáil’sto‘Mills’,butneitherof these partiesshowedany evidence of coordinationatlowerpreferences,andinneitherof themwas‘Hackett’consistently last.Overall, then, the datadosupporta readingof the vote as reflectingaleft-rightpolarity,since nearlyall of the membersof partiesdefiningthemselvesasleftist4 gave ‘Hackett’theirfirst preference, while nomemberof Fine Gael orFiannaFáil did – they,onaverage,ranked‘Hackett’just a bit betterthanfourth.It isimportanttoemphasise,however,thatthe successof ‘Hackett’ dependedonitsbeingranked higherthanfifthbyeventhese voters.We shouldnot,therefore,read the resultas a left-wing‘victory’butasindicatingthe centre of gravityinafieldwhere preferences were certainlyclustered,butstill widelydispersed. Some technical issues All votingsystemshave theoretical anomalies,inthe sense thattheirresultscandependonfactors that seemarbitrary accordingto democraticprinciples. One of the technical issueswiththe BCis that the resultof a BC countcan be affectedbywhetherornotcertain losing optionsare on the agendaat all,evenwhenthe voters’preferencesamongthe otheroptionsare unchanged.5 (Both PRSTV and first-past-the postcarrythe same risk.) Was the final BCresultdependenton a shortlistof five alternatives,ratherthan,say,fouror six?A close analysisof the results(notshownhere) indicates thateliminatinganyof the fourlosingalternativeswouldnothave affectedthe outcome.6 However,itistheoreticallypossible thatif the shortlisthadcontained asixthoption,there could have beena tie forfirstplace betweenHackettandMills usingthe basicBC method,anda victoryfor MillsusingMBC. ThisresultisshowninTable 5.7 Table 5. Possible result in six-option contest. See text for details. Since the BC isknownto be theoreticallyvulnerabletothiskindof result,we shouldnotbe too surprisedbythisexample. Whatisreassuringinthe presentcase isthata problematicresult can onlybe constructedbymeansof verystrictand unrealisticassumptions(see note 8foran explanation). 4 I.e. Eirígí, Labour, People Before Profit, Sinn Féin, United Left. 5 This is referred to in the academic literaturewith the unfortunate name of the principleof ‘independence of irrelevantalternatives’(IIA). 6 This was tested by computing the resultof every four-option casethat included Hackett. 7 Table 5 is based on stipulatingthateveryone who preferred Mills to Hackett also preferred the sixth option to Hackett, but everyone who preferred Hackett to Millsranked the sixth option last.Note that no changes have been made to voters’ preferences among the original fiveoptions.Itis an accidentof the set of preferences that the basic BC method generates a tie. If Hackett had won the original contestby a slightly lower margin, Mills would havewon the six-option contest under both methods. Bermingh Duff Hackett Mills Stoker AN Other first preferences 15 1 27 6 2 0 second preferences 11 1 5 25 6 2 third preferences 10 8 2 13 6 5 fourth preferences 6 13 4 3 11 5 fifth preferences 3 12 5 1 12 7 sixth preferences 0 6 7 0 4 32 BC score 209 112 224 224 127 91 MBC score 198 108 197 204 125 61
  5. 5. A relatedissue is thatall votingsystemsare theoreticallyopentomanipulationby ‘strategicvoting’, i.e. votingdifferentlyfromone’strue preferencesforthe sake of achievingamore favoured outcome. Forexample,infirst-past-the-postsystems,itmakesstrategicsensetovote forone of the twoleadingcandidates,evenif youwouldprefersomeone else. InaBC system, the apparently obviousstrategyisto place the alternative youthinkof asyour biggestthreat atthe bottomof your list, ,so as to maximise the gapbetween yourpreferredoptionandthatalternative.Inthe current example,proponentsof ‘Mills’mighthave done soby ranking‘Hackett’ fifth, regardlessof their actual preferences,whileproponentsof ‘Hackett’mighthave done the opposite. It ishard to know whetheranycouncillors attemptedtovote strategically,butthe distributionof preferencesdoesnotsuggestthatthiswasat all widespread.Forexample,therewasnoconsistent patterninthe lowerpreferencesof voterswiththe same firstpreference. We can, however, see what would have happened if everyone who ranked ‘Mills’ above ‘Hackett’ had bumped ‘Hackett’ down to fifth place, and vice versa. Table 6 shows the result. Table 6. Result of simple strategic voting by ‘Mills’ and ‘Hackett’ supporters. What isstrikinginthisresultisthat the strategywouldhave givenvictoryto ‘Bermingham’,an alternative thatmanyvotersconsideredworse thaneither‘Hackett’ or‘Mills’. Infact,the outcome of thiskindof strategicvotinginBC systemsis very unpredictable,because whenyoudropthe rank of one rival,youraise the scoresof others.Supportersof BC argue that thisunpredictability discouragesstrategicvotingaltogether. In a basicBC vote,anothersimpleformof strategicvotingisto refuse tovote for anyoptionyou wouldnotlike towin,since thismaximisesthe differencebetweenthe pointsassignedtoone’s favourite option(s) andthose others.Itispossible thatsome of the councillorswhoreturned incomplete ballotswereactingstrategicallyinthisway,buttheymayalsohave been simply indifferentamongthe remainingoptions.The factthat‘Hackett’wouldhave wonunderthe MBC indicatesthatthisstrategy,if thatis whatit was,wasnot decisive in the presentcase. The effectof thisstrategy isalso highly unpredictable:atitslimit(whereeveryvoterranksonlyone candidate),it isequivalenttofirst-past-the-postanddefeatsthe purpose of aBC procedure. One of the advantagesof the MBC isthat it eliminatesthisform of strategicvoting. Conclusion DublinCityCouncil’suse of the BordaCountto name itsnew bridge wasan important milestonein decision-making.A detailedanalysisof itsresultsinboththe short-listingprocessandthe final decisionshowsthatitperformeditspurpose effectively. Althoughtherewere significantparty- Bermingh Duff Hackett Mills Stoker first preferences 15 1 27 6 2 second preferences 22 3 5 9 8 third preferences 5 19 0 4 14 fourth preferences 3 20 0 1 18 fifth preferences 0 0 18 24 0 BC score 184 114 173 104 120 MBC score 175 111 145 98 120
  6. 6. political differencesinhowcouncillorsvoted,the BCsystemexhibitedone of itscentral virtues, namely itschoice of an optionthathad at leastsome supportevenamongitsopponents. The technical analysisof the vote givenabove indicatessome of the waysthat the outcomesof the BC, like those of othersystems, canbe affectedbyseeminglyinsignificantissues,suchaswhether or not a losingalternativeisonthe ballotandhow incomplete ballotsare treated.However, itseems clearthat these issuesdidnotaffectthe outcome onthisoccasion. Although,like all voting procedures,BCsystemsare inprinciple opentostrategicvoting,there isnoevidence inthe present case of people votingstrategically,andthe case illustratesthe unpredictableeffectsof attemptingto do so. At the endof the day,what mattersisthat a votingprocedure reflects,asfaras possible,whatitis tryingto measure.BordaCountproceduresattempttofindthe alternative (orsetof alternatives) that has the mostsupportfrom the electorate,takingtheirentire setof preferencesintoaccount. Choosingacommemorative name foranew bridge inthe centre of Dublinshouldsurelybe basedon that kindof information.Itistherefore bothappropriateandadmirable thatDublinCityCouncil used the Borda Countto make its decision. References DublinCityCouncil (2013a) 'MonthlyCityCouncil Meeting02/09/2013: Minutes',(Dublin:DublinCity Council). --- (2013b) 'Naming of New City Bridge: Report No. 272/2013', (Dublin: Dublin City Council). Emerson, Peter (2012) Defining Democracy: Voting Procedures in Decision-Making, Elections and Governance (2nd edn.; Heidelberg: Springer).