IMPROVING AMENDMENT

transcript

Improving Amendment

JonathanL.Marshfield∗

I. INTRODUCTION State constitutional amendment rules are often

criticizedfortheirpoordesign.1Themostcommoncriticismis that the frequent use of direct democracy bypasses thevirtues of representative decision making and effectivelysurrenders constitutional politics to well-financed specialinterests.2 There is much evidence to support this view.Researchsuggeststhatcitizensusuallygivelittlethoughttohow they vote on initiatives, rely on only one source ofinformation, and rarely discuss ballot initiates with morethanoneperson.3 Consequently,citizen-initiativeamendm-ents are often ill-considered, poorly vetted, and evendiscriminatory.4Theinitiativehasalsobeen“industrialized”in many states.5 Professional signature-gathering firmsexertsignificantinfluenceontheballotagenda,andspecial- ∗ AssistantProfessor,UniversityofArkansasSchoolofLaw.IamgratefultoTaylorBishforhelpfulresearchassistanceinpreparingthisessay.

1. See Thomas Gais & Gerald Benjamin, Public Discontent and the Decline ofDeliberation:ADilemmainStateConstitutionalReform,68TEMP.L.REV.1291,1291-92(1995).

2. Eighteenstatespermitcitizenstoamendthestateconstitutionbyusingthecitizeninitiative.SeeJohnDinan,StateConstitutionalDevelopmentsin2014,in47THEBOOKOF THE STATES: 2015EDITION3, 3 (2015). For ahelpful summaryof themanycriticisms of the citizen initiative to amend state constitutions, see Cody Hoesly,Comment,ReformingDirectDemocracy:Lessons fromOregon, 93CALIF.L.REV.1191,1202-12(2005).

3. SeeGais&Benjamin,supranote1,at1301(summarizingempiricalresearchregardingvoterdecision-makingonballotmeasures).

4. Seeid.at1301-02(describingshortcomingsoftheinitiativeprocess);Hoesly,supra note 2, at 1209-12 (describing “discriminatory capture” of constitutionalinitiative).

5. SeeHoesly,supranote2,at1202(describinghowthe initiativeprocesshasbeen commoditizedbyprivate initiative firms that exert significant influence in theprocess).

478 ARKANSAS LAW REVIEW [Vol. 69:477

interestfirmsoftenspendsignificantmoneyonmass-mediacampaigns that affect referenda outcomes.6 Thus, to theextent constitutional amendment should involve a moredeliberate and inclusive democratic process, the initiativeseemstobeperformingpoorlyinmanystates.7

One solution might be to give the legislature greatercontrol over the amendment process.8 But this approachcomes with its own costs. Because of self interest inretaining the political status quo, state legislators oftenresist popular constitutional reform on important issuessuch as legislative term limits, redistricting, and campaignfinance.9Inadditiontothesetroublingagencycosts,citizenshave very few opportunities for meaningful participationand deliberation when the legislature controls theamendmentprocess.Althoughlegislaturessubmitproposedamendments to a public vote,10 referenda are oftenineffective at fostering meaningful citizen deliberation andparticipation. In fact,votersoftenskipreferendaquestions

6. ReidWilson,InitiativeSpendingBoomsPast$1BillionasCorporationsSponsorTheirOwnProposals,WASH.POST(Nov.8,2013),https://www.washingtonpost.com/blogs/govbeat/wp/2013/11/08/initiative-spending-booms-past-1-billion-as-corporations-sponsor-their-own-proposals/ [https://perma.cc/7U96-98G6] (statingthat corporations spent “more than $1 billion” on ballot initiatives in eleven statesbetween2012-2013).

7. See Gais & Benjamin, supra note 1, at 1302 (“[I]nitiatives are much lesssuccessful in producing deliberate, comprehensive, and representative change.”)(emphasisomitted);seealsoDAVIDB.MAGLEBY,DIRECTLEGISLATION:VOTINGONBALLOTPROPOSITIONS IN THE UNITED STATES 198-99 (1984) (concluding that initiatives aredominatedbyspecial-interestgroups).

8. See Eric Lane,Men are not Angels: The Realpolitik of Direct Democracy andWhatWe Can do About It, 34WILLAMETTE L. REV. 579, 580-81 (1998) (arguing thatdirectdemocracyinanyformshouldbeopposedinstatesinfavorofrepresentativelawmakingprocesses).

9. See, e.g., Anne G. Campbell, Direct Democracy and Constitutional Reform:CampaignFinance Initiatives inColorado, in 1 STATECONSTITUTIONS FORTHETWENTY-FIRSTCENTURY175,191-92(G.AlanTarr&RobertF.Williamseds.,2006)(explainingthat Colorado General Assembly stalled campaign finance reform); see also Gais &Benjamin,supranote1,at1298(“Legislaturesresist fundamentalrevisionsbecausemanyof thedemands forreformare in factaimedatstate legislaturesandthreatenthe interests of their members.”) (emphasis omitted); Heather K. Gerken, GettingFrom Here to There in Redistricting Reform, 5 DUKE J. CONST. L. & PUB. POL’Y 1, 1-2(2010) (noting that political reform regarding state districting is unlikely to comefromstatelegislaturesbecause“foxesareguardingthehenhouse”).

10. This is true for all states exceptDelaware,which allows the legislature toadoptamendmentswithoutapublicreferendum.SeeDinan,supranote2,at4-5.

2016] IMPROVING AMENDMENT 479

entirely,11 and research on voter turnout suggests thatcandidate elections, rather than ballot issues, truly drivevoterturnout.12

It seems, therefore, thatmany states need to improvethe democratic quality of their amendment procedures.States need effective ways to foster constructive publicdeliberation, incentivize meaningful citizen participation,andprovidechecksontheinfluenceofspecialinterests.13Inthisessay, Iconsiderwhetherstatesmightachievesomeofthoseimprovementsiftheychangedtheprocessforratifyingcitizen-initiative amendments to require debate andapproval by locally elected governing bodies rather than apublic referendum.14 Sending amendment ratificationdecisionsto locallyelectedbodiescouldhavethebeneficialeffect of keeping constitutional decision-making close tocitizenswhileatthesametimeretainingmanyofthevirtuesassociated with representative decision-making. It mightalso help undermine special-interest capture by dividingamendmentpower acrossnumerous independently electedbodiesratherthancentralizingitwithinastatelegislatureorpopularmajorityvote.

To explore the costs and benefits of this proposal, Ifocusmyanalysisonthesixteenstates thatcurrentlyallow

11. SeeMartinP.Wattenbergetal.,HowVoting IsLikeTakinganSATTest:An

AnalysisofAmericanVoterRolloff,28AM.POL.Q.234,247-48(2000)(explainingthatvoterstendtoskipmanyballotquestions);seealsoMAGLEBY,supranote7,at105.

12. SeeGais&Benjamin,supranote1,at1302.13. Id. at1303 (“Whatweneed... are constitutional revisionprocedures that

are deliberative as well as legitimate—procedures that command legitimacy byproviding for direct citizen participation and control, but that also generate andassess alternativeproposals, take into account thebest available information abouttheir likelyeffects, consider the interactionsbetweentheproposedchangesandtherest of the constitutional structure, and afford opportunities for discussion andaccommodationamongsignificantpoliticalinterests.”).

14. AlthoughCongresshasrequirednewstatestoincludeareferendumaspartof their amendment processes, there is no legal requirement that admitted statesretain the referendum. Indeed, Delaware’s current amendment rules allow thelegislaturetoamendtheconstitutionwithoutareferendum,andseveralstateshaveutilizedamendmentproceduresinthepastthatdidnotrequireareferendum.JennieDrage Bowser, Constitutions: Amend with Care, ST. LEGISLATURES, Sept. 2015, at 16,http://www.ncsl.org/research/elections-and-campaigns/constitution-amend-with-care.aspx[https://perma.cc/KBM9-T9RL].

for constitutional amendment by both legislative referral(where the legislatureproposes amendments to be ratifiedby a referendum) and citizen-initiative (where citizens canbypass the legislature and propose amendments to beratified by a referendum).15 The specific proposal that Iexplore iswhetherthosesixteenstatescould improvetheiramendment processes by changing amendment rules torequireratificationofcitizen-initiativesbysomemajorityofexisting county governing bodies rather than a statewidereferendum.16

I conclude that a “county-ratification” model has thepotentialtoimproveamendmentprocessesinatleastthreeways. First, it could improve the quality of citizenparticipationby reducing the sizeof the jurisdictionwhereamendmentdecisionsaredebatedanddecided.17Second,itcould increase the quality of public deliberation regardingconstitutional changeby taking amendmentdecisions fromthe secrecy and isolation of the voting booth and placingthem in theproverbial “townsquare”where locallyelectedofficialsmustpubliclyjustifyanddebatetheirdecision.18 Acounty-ratificationmodelmightalso increase thequalityofpublic deliberation by ensuring thatminority communitieshave a voice in the process.19 Third, a county-ratification

15. Statesallowingthecitizeninitiativeinclude: Arizona,Arkansas,California,Colorado,Florida,Michigan,Mississippi,Missouri,Montana,Nebraska,Nevada,NorthDakota,Ohio,Oklahoma,Oregon, and SouthDakota. See Dinan, supra note 2, at 15(listingstateamendmentprocesses). AlthoughIllinoisandMassachusettsalsoallowforamendmentbyinitiative,Idonotincludetheminmyanalysisbecausebothstatesplace significant limitationson the citizen initiative. Illinois allowscitizen-initiativeamendments regarding changes to only the legislative article of the Illinoisconstitution. See ILL.CONST. art.XIV, §3. Massachusettshasan “indirect initiative”process,whereby the legislaturemustapproveall citizen initiativesbefore theycanbeplacedon theballot. SeeMASS. CONST. art. XLVIII, pt. 4, § 5. Massachusetts alsoallows a supermajority of the legislature to amend initiative proposals. See MASS.CONST.art.XLVIII,pt.4§3.Becauseoftheselimitations,theinitiativeisaninfrequentmethodofconstitutionalchangeinbothstates.Indeed,neitherstatehasadoptedanamendmentbyinitiativeinthelastdecade.

16. As explained in more detail below, all sixteen states already havefunctioningcountygoverningbodies thatcouldtakeupproposedamendments. Seeinfratextaccompanyingnote167.

17. SeeinfraSectionIV.C.1.18. SeeinfraSectionIV.C.2.19. Id.

model might limit the influence of special interests bydividing the amendment power between multiple countygoverningbodiesthathaveincentivestomonitoreachotherandaregenerallymoreresponsiveandaccountabletolocalconstituencies.20

Of course, the county-ratification model is not apanacea. There are many difficulties and costs associatedwiththisapproach.Itmight,forexample,makethecitizen-initiativetoodifficulttouse,whichwouldeffectivelyshiftallamendment power to the legislature.21 Countyrepresentativesmight alsobe ill-suited todecide statewideconstitutional issues because of mismatched expertise andlimited resources.22 A county-ratificationmodel could alsoresult in unconstitutional voter dilution because ofsignificantpopulationdifferencesbetweencounties.23Theseissues,amongothers,representseriousdifficultieswiththecounty-ratification model that cannot be overlooked. Mygoal in this essay is only to suggest that the county-ratificationmodel deserves serious consideration as statesstruggle with how they might improve their amendmentprocesses.

PartIIprovidesabriefsummaryofstateconstitutionalamendment rules and practices with an emphasis on thestates’ tradition of assessing and redesigning amendmentrules from time-to-time. Part III explores the majorproblemswiththetwodominantamendmentmethods:thelegislative-referral and citizen-initiative methods. Part IVpresentsmycounty-ratificationmodelandarguesthatithasthe potential to improve the democratic quality of stateamendment practices. Part V addresses some importantdifficulties and likely costs associated with the county-ratificationmodel.

20. SeeinfraSectionIV.C.3.21. SeeinfraSectionV.B.22. Id.23. SeeinfraSectionV.A.

II. STATE CONSTITUTIONAL AMENDMENT METHODS AND DESIGN

BecauseArticleVof theUnitedStatesConstitutionhasremained unchanged since 1788 and seems immune toredesign, it is easy to overlook the states’ long tradition ofexperimenting with amendment procedures.24 This Partprovides a very general overview of the dominantamendment methods that the states have considered anddevelopedovertime.Asothershavenoted,“nocatalogueofthe mechanisms for state constitutional change can fullycapture the richness or the variety of the approaches thathavebeenused.”25 Thus, thepurposeof thisPart isnot tochronicle all developments in state amendmentdesign, butonlytodemonstratethatthestateshaveastrongtraditionofreevaluatingand redesigning their amendmentprocedures.This Part also provides an original tabulation of recentamendment-ratedata,whichsuggestthatdespitevariety inamendment processes, the vast majority of stateamendments occur through two processes: (1) legislativeproposals ratified by referenda, and (2) citizen initiativesratifiedbyreferenda.

A. The State Tradition of Redesigning Amendment Processes

The whole idea of incremental constitutional changethrough textual amendment originated in stateconstitutions.26 Following John Locke’s 1669 FundamentalConstitutions of Carolina, which famously declared that

24. Seegenerally JOHN J.DINAN,THEAMERICANSTATECONSTITUTIONALTRADITION29-63 (2006) (describing themany revisions to state amendment procedures overtime).

25. G.AlanTarr&RobertF.Williams,GettingfromHeretoThere:Twenty-FirstCenturyMechanismsandOpportunitiesinStateConstitutionalReform,36RUTGERSL.J.1075,1077(2005).

26. See WILLI PAUL ADAMS, THE FIRST AMERICAN CONSTITUTIONS: REPUBLICANIDEOLOGYANDTHEMAKINGOFTHESTATECONSTITUTIONSINTHEREVOLUTIONARYERA136-42 (Rita Kimber & Robert Kimber trans., 1980) (discussing how early stateconstitutional theory broke from thinking of the time regarding supremacy andinalterabilityof constitutional law);GORDONS.WOOD,THECREATIONOFTHEAMERICANREPUBLIC:1776-1787,at614(1969).

“these fundamental constitutions... shall... remain...unalterable,”27thestatesbegantorecognizetheneedforanorderedprocessofconstitutionalchange.28

As the states began to experiment with how best todesignanamendmentprocess,theyweregenerallywearyofamendment procedures that empowered the legislature orother representatives to amend the constitution withoutdirect popular involvement.29 Only three of the twenty-sixstate constitutions adopted before 1800 authorized thelegislaturetoamendtheconstitution.30

The dominant early approach to amendment was torequireapopularlyelectedconstitutionalconventionforallamendments.31 Thoseconventionswereelectedseparatelyfrom the legislature and convened to consider specific,predetermined issues.32 This “convention model” ofamendmentwas intended to preserve popular sovereigntybyensuringthatelectedrepresentativescouldnotabusetheamendment power.33 Generally, conventions wereauthorized to adopt amendments without any furtherratificationorapopularreferendum.34Thedeliberationand

27. Sanford Levinson, Introduction: Imperfection and Amendability, in

RESPONDINGTOIMPERFECTION:THETHEORYANDPRACTICEOFCONSTITUTIONALAMENDMENT3, 4 (Sanford Levinson ed., 1995) (quoting and discussing the 1669 FundamentalConstitutionsofCarolina).

28. See G. ALAN TARR, UNDERSTANDING STATE CONSTITUTIONS 70-75 (1998)(discussingdevelopment of early state amendmentprocedures). Six of the originalsixteenstateconstitutionsdidnotcontainanyrulesforamendmentorrevision. Seeid.at62-63(listingearlystateswithoutamendmentrules).

29. See ADAMS, supra note 26, at 142 (“The point at issue clearly was therelationshipbetweenthesovereignpeopleandtheirelectedrulers.”).

30. See TARR, supra note28, at 61, 73-74 (summarizingdevelopment of earlystateamendmentprocedures).

31. SeeDINAN,supranote24,at41;TARR,supranote28,at73-74.32. See WALTER FAIRLEIGH DODD, THE REVISION AND AMENDMENT OF STATE

CONSTITUTIONS28-29(1910).33. SeeTARR,supranote28,at70-71.34. Id. The 1784 New Hampshire constitution was the first to adopt a

referendum requirement stating that all amendments adopted by the conventionmustbeapprovedbyamajorityofthe“qualifiedvoters.”DODD,supranote32;JamesA.Henretta,RethinkingtheStateConstitutionalTradition,22RUTGERSL.J.819,826-31(1991) (noting “[o]nly two of the twenty-eight state constitutions adopted before1800...weresenttothepeople forratification,”andthatamendmentrules“placedcontrolofthatprocessinthehandsofthelegislature”).

approval of a convention was considered sufficientdemocraticprocesstolegitimateamendments.35

By the mid-1800s, most states recognized that theconvention process was too inflexible.36 Rapidly changingcircumstances put pressure on state constitutional texts tochange,andtheconventionmodelwastoocumbersomeandrigid toaccommodate thoseneeds.37 Thus, statesbegan todebatealternativedesignsfortheamendmentpower.38Thedominant design concern during this period was to makeamendment easier and more responsive to changingcircumstances while retaining appropriate constitutionalstability.39

The states experimentedwith a variety of approachesthat authorized legislatures to adopt amendments.40 Somestates, for example, authorized the legislature to adoptamendments with a supermajority in two successivelegislativesessions.41 Thisdesignwasintendedtofacilitatedeliberation within the legislature and ensure publicaccountability with an intervening legislative election.42Other states required amendments to be separatelyapprovedbysupermajorities inbothhousesandsubmittedto a public referendum.43 As states debated and deployedthesevaryingdesigns,thedominantapproachthatemergedwas to require supermajority approval by the legislature,followedbyapopularreferendum.44

The next major development in state amendmentdesignwastheproductoftheProgressivemovementoftheearly twentieth century.45 As a result of state legislatures

35. TARR,supranote28,at70.36. DINAN,supranote24,at32-37,41.37. Id.38. Id.at41-47.39. Id.at42-44;DODD,supranote32,at120.40. DINAN, supra note 24, at 43-44; see also DODD, supra note 32, at 118-20

(recountingtheoriginsoflegislativeamendments).41. DINAN,supranote24,at43.42. Id.;DODD,supranote32,at122.43. DINAN,supranote24,at43.44. Id.at44-45.45. Id.at47-48.

and judges blocking popular Progressive social reformlegislation, there was a growing populist movement tofurther “liberalize” amendment procedures.46 The mostsignificant design change during this period was theadoption of the constitutional-initiative as an entirely newmechanism for amendment.47 Thegeneral structureof theconstitutional-initiativewastoallowcitizenstoproposeandratify constitutional amendmentswithout any involvementby the legislature.48 The initiativewas intendedtoaddressconcerns regarding special-interest capture in statelegislatures and facilitate greater citizen participation inconstitutionalpolitics.49Oregonwasthefirststatetoadoptthe constitutional-initiative in 1902, followed by seventeenotherstatesbytheendofthetwentiethcentury.50

The modern constitutional commission is a furtherdevelopment (or perhaps refinement) in state amendmentdesign.51 In general, commissions are independent bodiesthat provide recommendations for constitutional change tothe legislature, a constitutional convention, or the peopledirectly.52Commissionsaregenerallyestablishedtoprovideexpert consideration of constitutional reform, but they donothaveamandatetoadoptconstitutionalchangesontheirown.53 Florida’s Constitutional Revision Commission,constitu- tionalized in 1968, is the one exception to thisstructure.54 The Florida Commission is convened

46. Id.at47-50.47. Id.at59.48. DINAN,supranote24,at59.49. Id.at59-60.50. Id.at62,313n.132(listingtheeighteenstatesanddatesofadoption).51. SeegenerallyTarr&Williams,supranote25,at1094-1100(describingthe

variousformsofconstitutionalcommissions).52. See Robert F.Williams,Are State Constitutional Conventions Things of the

Past? The Increasing Role of the Constitutional Commission in State ConstitutionalChange,1HOFSTRAL.&POL’YSYMP.1,4(1996).

53. Id. at 5. Commissionmembers areusually appointed rather than elected,butappointeescanbealready-electedofficials,citizens,experts,stakeholders,oranycombination of these. Id. at 2 (characterizing state constitutional commissions as“appointed”).

54. See generally Tarr &Williams, supra note 25, at 1097-99 (describing theFloridacommissionprocess).

automaticallyeverytwentyyears,anditsrecommendationsare automatically placed on the ballot for ratification byreferendum.55 Florida’s variation on the constitutionalcommission is significant from the perspective ofconstitutionaldesignbecauseitwasintendedtooperateasacheck on legislative capture and to allow an independentbody to propose necessary amendments directly tocitizens.56

Current state amendment procedures representvariations on and combinations of these core components.Forty-nine states currently permit amendments bylegislative approval, followed by a public referendum forratification.57 Eighteen of those states also permitamendments by constitutional-initiative, although some ofthose states limit the use of the initiative to certainsubjects.58Delawareissomewhatuniqueinthatitdoesnotallow for amendment by public initiative and is the onlystate that permits amendment by the legislaturewithout aratifying referendum.59 At least three states still allow foramendments to bemadeby calling a limited constitutionalconvention to consider specific reforms.60 And, as notedabove, Florida’s amendment rules require a constitutionalcommissiontobeformedeverytwentyyearsandauthorizesthe commission to put proposed amendments directly to areferendum.61

A few states have experimented with integrating theconstitutional-initiative and legislative-referral methods ofamendment. Massachusetts, for example, requires that all

55. Id.at1097.56. Id.at1098n.122.57. SeeDinan,supranote2,at13tbl.1.2.58. See id. at 15 tbl.1.3; see also, e.g., MASS. CONST. art. XLVII, § 2 (excluding

certainsubjectsfromamendmentbyinitiative).59. See id. at 13-15 tbls.1.2 & 1.3. The Delaware Legislature must approve

amendmentsbytwo-thirdsmajoritiesintwosuccessivelegislativesessions.Id.at13tbl.1.2.

60. See Tarr & Williams, supra note 25, at 1086 (discussing Kansas, NorthCarolina,andTennessee).

61. Id. at 1097. Rhode Islandhas a hybrid approach that requires a periodiccommissiontoconsiderpresentingspecificconstitutionalquestionstotheelectorateandwhethertheelectorateshouldcallaconstitutionalconvention.Id.at1099-1100.

constitutional initiatives be approved by one-fourth of thelegislature sitting jointly in two successive legislativesessions.62 Similarly, Mississippi allows the legislature tosubmit an amended or alternative proposal to votersalongside any initiative proposals.63 These refinements onthe amendment process seem designed to provide a checkondirectdemocracybysubjectingtheinitiativetoreviewbyelectedrepresentatives.

In sum, the states have a tradition of revising andexperimentingwithamendmentprocesses. It is important,therefore, that scholars assess existing amendmentprocesses and provide useful guidance as to howamendmentprocessesmightbeimprovedinthefuture.

B. Contemporary State Practice of Formal Amendment Despite the various amendment methods available

understateconstitutions,64dataregardingtheactualuseofamendment procedures are helpful in focusingrecommendations for design reform. Two points areparticularly important for present purposes: (1) statesamend theirconstitutionsrelatively frequentlyonavarietyof important issues, and (2) despite the various stateamendment processes, 99.5% of amendments in the lastdecade occurred through either the legislative-referral orthecitizen-initiativemethod,65makingthesetwoapproachesthedominantmethodsofstateconstitutionalchange.

62. SeeDinan,supranote2,at15tbl.1.3.Massachusettsalsolimitedtheuseof

the initiative by excluding certain subjects from amendment by initiative. MASS.CONST.art.XLVII,§2.

63. Dinan,supranote2,at15tbl.1.3. Illinois limits theuseof the initiativetoonlyamendmentsregardingthestructureofthelegislativebranch.Id.

64. See Teresa Stanton Collett, Judicial Independence and Accountability in anAge of Unconstitutional Constitutional Amendments, 41 LOY. U. CHI. L.J. 327, 334-35(2010)(notingtherearecurrentlyfourdominantmethodsofamendment:“(1)voteradoption of legislatively-referred proposals, (2) voter adoption of citizen-initiatedproposals, (3) voter adoption of commission-referred proposals, or (4) throughconstitutionalconventions”)(footnotesomitted).

65. Seeinfratextaccompanyingnotes71-73.

First, as many others have noted, state constitutionalamendment has become very frequent in most states.66Although there is great variety, states revise theirconstitutions every three years on average.67 Additionally,those amendments touch on a variety of importantstructuralandrightsissues.68Amendmentpracticeisthusasignificant aspect of constitutional practice in the states.Unlike the United States Constitution,where constitutionalchange occurs primarily through judicial review of aninfrequently amended text, state constitutional change ismuch more institutionally interactive, with democraticinstitutionsplayingasignificant role in thedevelopmentofconstitutional rules.69 It is therefore important thatamendment processes be critically examined and refinedwhennecessary.

Another important observation from state amendmentdata relates to the relative use of the various methods ofamendment. Althoughthestateshaveexperimentedwithavariety of different amendment approaches, amendmentsseem to occur primarily through legislative referral toreferenda or citizen initiatives.70 There have been a fewamendmentsbytheconventionorcommissionmethods,butjust over 91% of the 575 state constitutional amendmentsadopted over the last decade occurred through the

66. SeeMila Versteeg & Emily Zackin,American Constitutional Exceptionalism

Revisited,81U.CHI.L.REV.1641,1672-75(2014).67. Id.at1674.68. There were 603 proposed amendments to state constitutions (excluding

local amendments) between 2002 and 2008. See John Dinan, State ConstitutionalDevelopments in2008, in41THEBOOKOFTHESTATES:2009EDITION3,6tbl.B(2009);JohnDinan,StateConstitutionalDevelopmentsin2007,in40THEBOOKOFTHESTATES:2008 EDITION 3, 5 tbl.B (2008). Of these, 242 related to issues of governmentstructure, and 110 related to rights issues. This means that more than fifty-eightpercent of all proposed amendments to during this period concerned issues ofgovernment structure or rights. Of the proposed amendments that were actuallyadopted,anevenhigherpercentagerelatedtostructuralorrightsissues.

69. Richard Albert, American Exceptionalism in Constitutional Amendment, 69ARK.L.REV.217(2016).

70. TARR,supranote28,at139;Dinan,supranote2,at4tbl.A.

legislative referral and referendum method.71Approximately 7.5% of amendments in the last decadeoccurred through citizen initiatives,72 and less than 1%occurredthroughcommissions.73

The data are interesting on a state-by-state level forstateswithatleasttwomethodsofamendment.74Thechartbelow illustrates the relative use of the methods ofamendmentinthesestates.75

71. Thispercentageiscalculatedbasedonmyoriginaltabulationofamendment

data as reported in the annual Book of the States from 2005-2015. Of the 575amendmentsadopted,527ofthemwerethroughthelegislative-referralmethod.

72. Forty-five of the 575 amendments were adopted through the initiativeprocess.

73. Onlythreeof the575amendmentswereadoptedthroughthecommissionprocess.

74. I did not include Illinois or Massachusetts in this chart because of thesignificantlimitationsontheinitiativediscussed.Supranote15.

75. ThesedataarefrommytabulationofamendmentdatafromtheBookoftheStates.Seesupranote71(explainingdatatabulation).Alltabulationsareonfilewiththeauthorandareavailableuponrequest.

0510152025

StateAmendmentsbyMethodofAmendment2005-2014

Legisla-ve Ini-a-ve Commission

As the data show, the legislative-referral methodremainsrelevanteveninstateswherecitizensareactive inusing the initiative. In California, for example, thirty-eightpercent of all amendments over the last decade wereinitiated by the legislature.76 Oklahoma is also interestingbecause it has adopted the most amendments in the lastdecadeofanyofthesestates(twenty-fiveamendments),butall of those amendments were by legislative referral.77Finally,itisinterestingtonotethattheaverageamendmentratebetweenstateswiththeinitiativeandstateswithouttheinitiative is very similar. The average amendment rate forstates without the initiative is 11.5 amendments over thelast tenyears.78 For stateswithboth the initiativeand thelegislative-referral process, the average amendment rate is11.3.79 Thus, it appears the contemporary amendmentpracticeinthestatesisdominatedbythesetwomethods.Itis now appropriate to consider existing criticisms of thesemethodsandhowtheymightbeimproved.

III. PREVAILING CRITICISMS OF STATE AMENDMENT METHODS

ThisPartprovidesabriefsummaryoftheprevailingcriticisms of the two dominant methods of stateconstitutional amendment: the legislative referral and thecitizeninitiative.

A. Criticisms of the Legislative-Referral Method Asnotedabove,thelegislative-referralmethodisthe

dominantmethodofamendmentinmoststates. Eachstatehas its own important history with adopting and

76. ThispercentageiscalculatedbasedonmyoriginaltabulationofamendmentdataasreportedintheannualBookoftheStatesfrom2005-2015.

77. ThispercentageiscalculatedbasedonmyoriginaltabulationofamendmentdataasreportedintheannualBookoftheStatesfrom2005-2015.

78. This includesMassachusetts and Illinois. The average amendment rate issimplythetotalnumberofamendmentsbetween2004and2015,dividedbythetotalnumberofstates.

79. Inthepastdecade,therewereonlythreeamendmentsadoptedbasedonacommission’sreferral.ThosewereinFloridain2008.

implementing this approach, but the legislative-referralmethod is generally intended to ensure that proposedamendments are deliberated by representatives withincentives to “foster compromise, continuity, andconsensus”80 before submitting vetted proposals to apopular vote. Notwithstanding its dominance in currentstate constitutional amendment practice, there are realproblems with the legislative-referral method ofamendment. This Section briefly describes the method’sgeneralshortcomings.

1. Legislative Self-Interest As Undermining Democratic Process

Thelegislative-referralmethodgivesstatelegislatorsthepowerto initiateamendmentsasrepresentativesof thepeople.81 In thiscapacity, legislatorsareexpected toact inthe interests of their constituents and the polity as wholerather than their own self-interest.82 However, manyconstitutional reforms create a significant conflict betweenlegislator-andconstituent-intereststhatcanunderminethedemocraticqualityofthelegislative-referralmethod.83

This is primarily because elected legislators have avestedpersonalinterestinretainingthepoliticalstatusquothatresultedintheirelection.84Constitutionalamendmentsrelated to issues such as redistricting, campaign finance,legislative term limits, and voter registration andidentificationoftencreateanincongruencebetweenpopularcollectivepreferencesandrepresentativedecisionmaking.85

80. DavidB.Magleby,LettheVotersDecide?AnAssessmentoftheInitiativeandReferendumProcess,66U.COLO.L.REV.13,43(1995).

81. SeeMatthewRobinson, Note,Deferring to Congressional Interpretations ofAmbiguousStatutoryProvisions,16N.Y.U.J.LEGIS.&PUB.POL’Y565,569(2013).

82. Seeid.at590.83. Gais&Benjamin,supranote1,at1298;seealsoGerken,supranote9.84. See Gerald Benjamin, Constitutional Amendment and Revision, in 3 STATE

CONSTITUTIONS FOR THE TWENTY-FIRST CENTURY: THEAGENDA OF STATE CONSTITUTIONALREFORM 177, 178 (G. Alan Tarr & Robert F. Williams eds., 2006); see also Gais &Benjamin,supranote1,at1298;Gerken,supranote9,at1-2.

85. See Campbell, supra note 9, at 176, 179-93 (discussing the legislature’sconflictof interestregardingcampaign-financereform);Gerken,supranote9 (same

This can result in representatives using the amendmentpower to serve their own interests by introducingamendmentsthatwouldpreservethepoliticalstatusquoorresistingamendmentsthatwouldchangethestatusquo.86

Thus, a major issue surrounding the legislative-referral method is the legislature’s ability to misuse theamendment power inways that further the self-interest ofindividual legislators or political parties and fail to respectthepeople’spreferences. Improvingthedemocraticqualityof the amendment process would suggest the need for amore direct method of amendment that is not entirelydependentonthelegislature.

2. Capture and Agency Costs As Undermining Popular Sovereignty

A related problem with the legislative-referralmethod is itsvulnerability to“capture”byspecial interests.Capture occurs when groups or people with a personalinterest in lawmaking gain a disproportionate influence in

regarding redistricting reform); David Orentlicher, Conflicts of Interest and theConstitution, 59WASH.&LEEL.REV.713,763-64 (2002) (sameregarding legislativeterm limits); Nicholas Stephanopoulos, Reforming Redistricting: Why PopularInitiatives to Establish Redistricting Commissions Succeed or Fail, 23 J.L. & POL. 331,331-32 (2007) (same regarding redistricting); Mark Thomas Quinlivan, Comment,OnePerson,OneVoteRevisited:TheImpendingNecessityofJudicialInterventionintheRealm of Voter Registration, 137 U. PA. L. REV. 2361, 2386 (1989) (same regardingvoter-registrationlaws). Voter-identificationlawsareaninterestingexampleofthisphenomenon because they can present a clear conflict of interest between citizensandrepresentatives,butrecentlytheyhaveinvolvedaction(ratherthaninaction)bystate legislatures to preserve the status quo. See Bob von Sternberg & Rachel E.Stassen-Berger,GOPProposingAmendmentstoAvertPossibleVetoonLegislation,STARTRIBUNE (Apr. 27, 2011, 11:05 PM), http://www.startribune.com/gop-proposing-amendments-to-avert-possible-veto-on-legislation/120831589/ [https://perma.cc/6PQA-DHCL] (discussing a proposed Minnesota amendment to require voteridentification as a way to entrench Republican political power). Interestingly, theamendmentwas placed on the ballot inMinnesota, but rejected by voters. MartinMoylan et al., Voter ID amendment defeated, MPR NEWS (Nov. 7, 2012, 1:46 AM),http://www.mprnews.org/story/2012/11/06/politics/voter-id-amendment [https://perma.cc/QG7G-GY5L].

86. Gerken,supranote9.

thelawmakingprocess.87 Whenthisoccurs,thelawmakingprocess is used in amanner that benefits a small group ofinterested stakeholders and is incongruent with popularpoliticalpreferences.88

Unfortunately, state legislatures have a history ofbeing captured (or “corrupted”) by special interests fromtime to time.89 If the legislature also controls theamendment power through the legislative-referralmethod,specialinterestscanentrenchfavorablepoliciesinthestateconstitutioneventhoughthepublicdoesnotsupportthem.90Additionally, if special interests control the legislature, thepeople have limited means of taking back control of theirconstitution.91 Thus, from a design perspective, if theamendment power is to provide a check on electedrepresentatives,itmustbeavailabletothepublicinamoredirectmannerthanthatprovidedbythe legislative-referralmethod.92

3. Limited Citizen Deliberation or Involvement Afurtherlimitationonthelegislative-referralmethod

is that it provides very few opportunities for meaningful

87. SeeEdward J. Janger, PredictingWhen the Uniform Law ProcessWill Fail:

Article 9, Capture, and the Race to the Bottom, 83 IOWA L. REV. 569, 632 (1998)(providingageneraldefinitionofgovernmentcapture).

88. Id.at584-85.89. TARR, supra note 28, at 110-13 (describing historical corruption of state

legislatures by special interest); see also State v. Miller, 45 Mo. 495, 498 (1870)(claimingthat therewasa“mostviciousandcorruptsystemwhichprevailed inourlegislative bodies”). The influence of special interests on state legislatures persiststoday. See EditorialBoard,Post’s Endorsement: VoteNo onBallotQuestion 1 inVa.,WASH. POST (Nov. 1, 2012), https://www.washingtonpost.com/opinions/posts-endorsement-vote-no-on-ballot-question-1-in-va/2012/11/01/62636726-1c8e-11e2-ba31-3083ca97c314_story.html [https://perma.cc/6VCK-PRGV] (describingeminent-domain amendment to Virginia constitution that was the product oflegislators“inthralltospecialinterests”).

90. SeeGais&Benjamin,supranote1,at1298,1314.91. SeeOraFredHarris,Jr.,ComplexProductDesignLitigation:ANeedforMore

Capable Fact-Finders, 79 KY. L.J. 477, 486 n.52 (1991) (“Because of politicalcompromising with and influence peddling by special interest groups, the stateconstitutionalamendmentprocessisparticularlyvulnerabletodilution.”).

92. See DINAN supra note 24, at 47-51, 53 (describing this debate in variousstateconstitutionalconventions).

citizen involvement in the amendment process.93 Thelegislative-referralmethodpermitscitizeninvolvementonlyattheratificationstage,andthenonlybycastingan“up-or-down” vote on the legislature’s proposal.94 Because of thesize and geographical dispersity of state populations,mostcitizens find state government too remote for them to getinvolved in “thick” forms of citizen participation, such asattending legislative debates and contacting staterepresentativesregardingproposedamendments.95

Additionally, even “thin” forms of participation, likevoting on amendment referenda, seem very poormethodsfor fostering citizen involvement and participation.Empirical research into voter choice suggests that citizensare generally more attentive to candidate elections at thestate level thanballotquestions,and thatmanyvotersskipballotquestionsaltogether.96

Thus, the legislative-referral method appears toprovidefewopportunitiesfordirectcitizeninvolvementanddeliberationregardingproposedconstitutionalchanges.

B. Criticisms of the Citizen-Initiative Method The citizen-initiative method of state constitutional

amendment is a relatively recent innovation. As notedabove,Oregonwasthefirststatetoadopttheconstitutional-initiativeprocess in1902.97 Therearenoweighteenstates

93. Seeid.at52(discussingneedforgreatercitizeninvolvementinamendmentprocess).

94. See Tarr & Williams, supra note 25, at 1092-94 (describing legislativereferralmethodofamendment).

95. Ihavearguedelsewherethatcitizensaremorelikelytoengageinthistypeof political activity at the state level than at the federal level. See Jonathan L.Marshfield,ModelsofSubnationalConstitutionalism,115PENNST.L.REV.1151,1187-88(2011).However,empiricalresearchsuggeststhatevenwithinstatejurisdictions,most citizens are unlikely to participate in this way when jurisdictions have morethan1millionpeople. See J.ERICOLIVER,DEMOCRACY INSUBURBIA42(2001)(findingthatonlytwenty-fivepercentofcitizensincitiesofonemillionpeopleormorecontactlocalrepresentatives).

96. SeegenerallyMAGLEBY,supranote7,at77-99(comparingvoter turnout inballot-question-onlyelectionstocandidateelections);Gais&Benjamin,supranote1,at1302(summarizingempiricalresearchregardingvoterdecision-makingonballotmeasures).

97. Supranote50andaccompanyingtext.

that permit citizens to amend their constitutions to somedegree through thismethod.98 Although each state has itsown history in adopting the constitutional initiative, statesgenerallyadoptedtheinitiativeasamechanismforrestoringpopularsovereigntytoconstitutional law.99 However,aftermore than a century of use in several states, significantproblemshavedeveloped in theuseof thecitizen-initiativemethod.

1. Circumventing Deliberation and Lawmaking Expertise Perhapsthemostcommonandpersistentcriticismof

the citizen-initiative method of state constitutionalamendment is that it circumvents constructive politicaldeliberation regarding fundamental political issues. Thiscritique echoes Madison’s famous argument in favor ofrepresentative decision making.100 Madison argued thatdirectdemocracycanresultinnear-sighteddecisionmakingvoid of administrative expertise and lacking relevantinformation.101 Madison also feared that direct democracywas more likely to be misused by interest groups infurtherance of goals that were not in the best interest ofsocietyasawhole.102 ForMadison,representativedecisionmakingwaspreferablebecauseitinvolvedinformeddebateby professional lawmakers who were best situated toevaluateaproposal’sbroadimpactonsocietyasawhole.103

From a design perspective, this process can befosteredbysubmittingproposalstoanopendeliberationbyelectedrepresentatives.AsHansLindenoted,“Deliberationin representative bodies does not often achieve its ideal of

98. See supra note 15 and accompanying text (discussing these states andlimitationsoninitiative).

99. SeeDINANsupranote24,at60(discussinghistoricalreasonsforadoptionofthe initiative); Magleby, supra note 80 (stating the initiative was a check on statelegislaturesandcourtsandservedtopromote“widespreadparticipation,openaccesstothepoliticalagenda,andpoliticalequality”).

100. SeeTHEFEDERALISTNO.10(JamesMadison).101. Id.102. Id.; see Hans A. Linde, When Is Initiative Lawmaking Not “Republican

Government”?,17HASTINGSCONST.L.Q.159,166(1989).103. SeeTHEFEDERALISTNO.10,supranote100.

dispassionate debate and logical persuasion, but it doesinstitutionalize” decision making in a way that facilitatesdiscussion regarding society’s interests as a whole ratherthan the particular preferences of individual citizens orgroups.104 Representatives are publically accountable tobroad-basedconstituenciesthatlikelyhavevariedinterests.They also deal repeatedly with fellow representativesregardingmanydifferentissues.Thus,representativesmustevaluatetheirvotesinabroadercontext.Theymustidentifyand rank public priorities, and make decisions about howbest to allocate limitedpublic resources as awhole. All ofthese factors can facilitate a more holistic and informeddecision-makingprocess.

Manycommentatorshavearguedthatamendmentbyinitiative circumvents this deliberative process.105 Thismethodallowsanycitizentobypassrepresentativedecisionmakingandplaceconstitutionalchangesontheballotsimplyby obtaining a set number of signatures from fellowcitizens.106Citizensthenvoteontheproposalbycastinganindependent and private ballot.107 Although this processbringsdecisionmakingdirectlytoindividualcitizens,itdoesnot provide a forum for deliberation between decision-makers and it assumes that citizens will independentlygather relevant information and obtain the expertisenecessary to decide an issue.108 The secrecy of the voterratification process also makes it susceptible to short-sighted,narrow,anddiscriminatorychoice.

Empirical research into the initiative processsuggests that these criticisms are well founded.109 Voterchoiceoninitiativesseemstobeheavilyaffectedbygeneral

104. Linde,supranote102,at169.105. See, e.g., Gais & Benjamin, supra note 1, at 1301-02 (describing

shortcomingsofinitiativeprocess).106. SeeM.DANEWATERS,THEINITIATIVEANDREFERENDUMALMANAC2-29(2003)

(describinginitiativeprocessinallinitiativestates).107. ARK.CONST.art.V,§1.108. Gais&Benjamin,supranote1,at1292.109. Id. at1301-02(“[R]esearchoncitizen initiativessuggest that theyhardly

resembleadeliberativeprocess.”).

discontent with bureaucracy and government officials’perceived inefficiencies rather than the specificmerits of aparticular proposal.110 Thismaybe in part because voterssuccumb to significant information difficulties that impedecritical decision making and constructive deliberation.111Onestudyfoundthatmanyinitiativevoterslearnabouttheinitiativeonlydaysbeforethevote,consultonlyonesourceof information regarding the initiative, and very rarelydiscuss the initiative with anyone.112 Initiative campaignmaterialalsosuggests“littleevidenceofreallearningabouttheissues.”113

These findingssuggestsignificantproblemswith thecontemporary constitutional-initiative. The initiativeprocess seems generally ineffective at producing informed,high-quality collective decisions that serve the interests ofsocietyasawhole.

2. Capture by Well-Funded Special Interests The citizen initiative was initially envisioned as a

populist check on government.114 Ideally, citizens wouldmobilize and use the initiative to correct governmentcorruption or capture and restore power to the people.However, a recurring criticism of the contemporaryinitiativeprocessisthatitishasbecomemoresusceptibletocapturebyspecialinterestgroupsthanstatelegislatures.115

There is compelling empirical evidence that specialinterests often control the initiative process.116 A niche

110. Id.at1301.111. Id.112. Seeid. Thismeansthatmanyinitiativesaredecidedbasedonoff-the-cuff

visceral reactions by voters rather than sustained deliberation between decisionmakers.

113. SeeGais&Benjamin,supranote1,at1302.114. SeeHoesly,supranote2,at1193(“Originally,backersofdirectdemocracy

believed that it would be ‘themedicine of the constitution’....”) (quoting JAMES D.BARNETT, THE OPERATION OF THE INITIATIVE REFERENDUM, AND RECALL IN OREGON 78(1915)).

115. Seeid.at1202.116. A compelling anecdotal example is a recent amendment to the Ohio

constitutionthatwouldhavelegalizedmarijuanaandgrantedaproductionmonopolytoanysupporteroftheinitiativethatcontributedatleast$2milliontothecampaign.

industry has developed around the initiative in manystates.117 Initiative firms will guarantee the requirednumber of signatures for an initiative “on a money backbasis,”andpromise tocollectvalidsignatures inas littleasforty-five days.118 The going rate for a guaranteed ballotinitiative in California, for example, was approximately $1millionin1998.119

Ballot industrialization has resulted in a massiveinflux of special-interest financing, making it increasinglydifficultforvolunteer-onlyinitiativestosucceed.120 Indeed,betweenMay2012andNovember2013,corporationsspentmore than $1 billion in ballot initiatives in only elevenstates,121 and the number of ballot initiatives using onlycitizenvolunteers is very small compared to initiatives runby firms.122 InOregon, for example, ninety-fourpercentofinitiatives on the ballot between 1996 and 2002 werehandled by paid signature-gathering firms, and eightypercent of initiatives on the ballot inWashington between1992 and 2000 were by paid gathering firms.123 In thewords of a signature-gathering firm inWashington, “thereisn’tachanceintheworldavolunteereffortisgoingtomake

SeeNoahFeldman,OhioRejectsPot,but itsConstitutionGetsWeird,BLOOMBERGVIEW(Nov.4,2015,1:32PM),http://www.bloombergview.com/articles/2015-11-04/ohio-rejects-pot-but-its-constitution-gets-weird [https://perma.cc/9RCB-35SY]. Anotherparticularly egregious example was Proposal 6 in Michigan regarding theconstruction of a new bridge into Canada. SeeMichigan Proposal 6 Results: VotersRejectBallotProposalRequiringPublicVoteonInternationalBridges,HUFFINGTONPOST(Nov. 7, 2012, 1:03 AM), http://www.huffingtonpost.com/2012/11/07/michigan-proposal-6-results-bridge_n_2084608.html[https://perma.cc/VB5F-ZMPY].Proposal6 would have amended the Michigan constitution to require a public referendumbeforethestatecouldspendanymoneytoconstructanynewinternationalbridges.Id. The proposalwas sponsored almost entirely by the owner of the only existingbridgebetweenDetroitandCanada.Id.

117. SeeHoesly,supranote2,at1202(describingthis“industrialization”oftheinitiativeprocess).

118. Seeid.119. Id.at1202-03.120. Id.121. SeeWilson,supranote6.122. SeeHoesly,supranote2,at1203.123. Id.; see also Richard J. Ellis, Signature Gathering in the Initiative Process:

HowDemocraticIsIt?,64MONT.L.REV.35,53,55(2003).

it.”124Additionally,specialinterestsspendsignificantcapitalsuccessfully opposing volunteer amendments that mightnegativelyaffectthem.125Onestudyfoundthatthisstrategyis particularly effective: “[O]pponents of an initiative win80%ofthetimewhentheyoutspendproponents.”126

Consequently, the initiative process is oftendominated by well-funded special interest groups ratherthan genuine populist movements.127 This has led someobservers to conclude that although the initiative wasintended to foster “grassroots democracy,” it has insteadbecome a haven for “greenback democracy.”128 Because“anybody can buy theirway on the ballot,”129 the initiativeoften “[a]llow[s] rich individuals or well-financed specialintereststoqualifymeasuresfortheballotalmostregardlessofeitherthedepthorintensityofpopularsupport.”130

There have been efforts to reform the initiativeprocessbyimposingrestrictionsonpaidsignaturegatherersand other requirements for signature gathering.131 Thesemeasures have had some impact, but they have not fullyaddressed theproblem,and theSupremeCourthas limitedthe type of restrictions states can place on signaturegathering.132 Indeed, most states with the initiative stillallow paid signature gatherers without significantlimitations,andtheprocessstillseemstooeasilydominatedbywell-fundedinterestgroups.

124. Ellis,supranote123,at54.125. SeeHoesly,supranote2,at1204-05.126. Id.at1205.127. See id. at 1202 (“Professional signature gatherers and political strategy

firms have long dominated the process.”); see also id. at 1206-09 (noting that thisproblem involves more than corporate interest groups; it also involves ideologueswithfundingfromnationalsources). ButseeEllis,supranote123,at54(discussingunfundedvolunteeramendmentsthataresuccessful).

128. SeeEllis,supranote123,at58.129. Seeid.130. Seeid.131. SeeHoesly,supranote2,at1212-13(discussingreformefforts).132. See Meyer v. Grant, 486 U.S. 414, 416, 428 (1988) (holding that states

cannotcriminalizepaidsignaturegathering).

3. Exasperating the Problem of Abusive Majorities Another criticism of the initiative is that it

exasperates majoritarian abuses. This critique isparticularly relevant to the use of the initiative in theconstitutionalcontext.

Contemporary constitutional theory posits that onefunction of a constitution is to protect minorities frommajorities who might use democratic institutions in adiscriminatorymanner.133Todothis,constitutionsmustbeentrenchedbeyondtherealmofordinarypoliticsandshouldbe subject to lawmaking processes that protect againstsimple aggregation of majoritarian preferences. Ifconstitutions are easily amended by a private vote of themajority, theyaremore likely tobeusedas instrumentsofdiscriminationratherthanminorityprotection.134

Criticsoftheinitiativeprocessnotethattheinitiativeisparticularlyvulnerabletomajoritarianabusesforatleastthree related reasons. First, the initiative bypasses anymeaningful public deliberation regarding constitutionalchanges thatmightaffectminorities.135 Publicdeliberationby elected officials can foster understanding and tolerancefor the needs and vulnerabilities of minorities and exposeprivateprejudicesthathavenoplaceinpubliclaw.136Tobesure, representative bodies are capable of discriminatoryaction, but transparent public deliberation can sometimesmitigateanimusbymakingitpublicandrequiringreasonedjustifications.

Second, the initiative process allows forconstitutional change by simple aggregation of confidentialcitizen preferences.137 An amendment is ratified when amajorityofcitizens,actingaloneandinprivate,registertheir

133. TerranceSandalow,JudicialProtectionofMinorities,75MICH.L.REV.1162,1179-80(1977).

134. This,ofcourse,isasimplificationofmanydeepandcomplicatedissuesinconstitutionaltheory.

135. Hoesly,supranote2,at1203.136. See JeremyWaldron,RepresentativeLawmaking,89B.U.L.REV.335,343-

45(2009).137. DINAN,supranote24,59-60.

supportfortheamendment.138Thisprocessbypassesmanyof the procedural safeguards that protect minorities fromabuses by other democratic institutions. Legislators, forexample, are subject topublic transparency for theirvotes.They cannot privately vote on explicitly discriminatorylegislation to pursue their own animus without publicexposure. Additionally, legislators must interact withconstituentsandotherlegislatorsonarecurringbasis.Theirvotingtrackrecordcanimpacttheirrelationshipswiththeirconstituents and colleagues, and, consequently, can affecttheir voting behavior on discriminatory legislation. In areferendum, on the other hand, proposals are decidedentirely on the confidential preferences of individualvoters.139 This suggests that the initiative is moresusceptible todiscriminatoryusesbecausevotersneednotdiscussorevenrevealtheirdecisionwithanyone.Empiricalstudies confirm that the initiative is often used to enactdiscriminatorymeasures.140

Third, referenda votes on initiatives seem tomarginalize minority voters. Empirical studies show that“[v]oting on ballot propositions only amplifies the socialclassbiasinparticipation,becausethosewithlowerincomesor lesseducation tend to skipvotingonballotquestionsatmuchhigher rates.”141 The referendumalso seems tobe aparticularly poormethod of includingminority viewpointsbecausethesimpleaggregationofvotesmeansthatminorityvotingblocswillbecanceledoutbymajoritieswithoutanyformal opportunities for discussion or debate betweenmajorityandminoritygroups.Consequently,“initiativesareoften decided by a minority of voters whose preferencesfrequentlydifferfromstatecitizensasawhole.”142

138. Magleby,supranote80,at13.139. AvigailEisenberg,When(ifEver)AreReferendumsonMinorityRightsFair?,

inREPRESENTATIONANDDEMOCRATICTHEORY3,8(DavidLaycocked.,2004).140. SeeHoesly,supranote2,at1209-12(listingexamples).141. SeeMagleby,supranote80,at33-34.142. SeeGais&Benjamin,supranote1,at1302.

IV. WHY INCLUDING COUNTY LEGISLATIVE BODIES MIGHT IMPROVE THE AMENDMENT PROCESSES

Despite the serious problems with existingamendmentprocedures,therehasbeenverylittleattentiongiventocreativealternatives.InthisPart,Iexplorewhetherincluding county legislative bodies in the amendmentprocess could improve the democratic quality ofamendment. I first describe the basic structure of countygovernment. I then provide a sketch of my proposal forincludingcountylegislativebodiesintheinitiativeprocess.Iconcludebyexplainingwhymyproposalmightimprovethedemocraticqualityoftheinitiativeprocess.

A. The General Structure and Function of County Governing Bodies

Counties are an intermediate level of governmentbetween state government and municipal/citygovernment.143 With the exceptions of Connecticut, RhodeIsland, and Virginia, every state establishes functioningcounty governments that collectively cover the entire stateterritory.144 The primary role of county government is toprovide basic government services to citizens within thecounty.145 Counties are also an important administrativearmof thestatebecausetheybridgethegapbetweenstatepolicy and local service delivery.146 County governmentshave also taken on an increasingly important role in thedeliveryandadministrationoffederalservices.147

143. SeeNAT’L ASS’N OF COUNTIES, COUNTY GOVERNMENT STRUCTURE: A STATE BY

STATE REPORT 6-8 (2009); OSBORNE M. REYNOLDS, JR., LOCAL GOVERNMENT LAW 24-25(4thed.2015).

144. SeeREYNOLDS,supranote143.145. Id.at25.146. See NAT’LASS’N OF COUNTIES, supra note 143, at 6; Tanis J. Salant,County

Governments: An Overview, INTERGOVERNMENTAL PERSPECTIVE, Winter 1991, at 7(providingahelpfulsummaryofcountygovernmentfunctions).

147. See Linda Lobao & David S. Kraybill, The Emerging Roles of CountyGovernments in Metropolitan and Nonmetropolitan Areas: Findings From a NationalSurvey,19ECON.DEV.Q.245,246(2005).

Countygovernmentsgenerallyhavejurisdictionoverincorporatedmunicipalitieswithintheirterritories,althoughstate law often grants incorporated municipalitiesindependence from counties regarding certain policies andservices.148 In large metropolitan areas, city governmentswill often “consolidate” with county government to createone body that governs the entire metropolitan area.149Despite these jurisdictional complexities, all citizens livingwithin states other than Connecticut, Rhode Island, andVirginia have representation in some kind of countygovernance, even if only within a consolidated city-countystructure.150

Thestructureofcountygovernmentvariesfromstatetostate,andinafewinstances,withinastate.151 Themostcommonstructure is the “commission form,”whichusuallyconsists of a governing body of between three and fiveelected representatives.152 These bodies often haveexecutive and legislative authority, and they sometimesshare responsibility for county administration with otherconstitutionally protected county officials such as countyclerks, auditors, sheriffs, and treasurers called “rowofficers.”153

Variations incountygovernance focusmostlyonthestructure of executive authority.154 Some states, such asArkansas and Florida, vest executive authority in a countyadministrator,separatefrombutappointedbythegoverningbody.155 Other counties have separately elected

148. See REYNOLDS, supra note 143, at 75 (Virginia is only exception to this;counties have no jurisdiction over territory covered by incorporated areas inVirginia).

149. SeeNAT’LASS’NOFCOUNTIES,supranote143,at8;seealsoJ.EdwinBenton,AnAssessmentofResearchonAmericanCounties,65PUB.ADMIN.REV.462,462(2005)(describingresearchintocountyconsolidation).

150. SeeREYNOLDS,supranote143andaccompanyingtext.151. SeeNAT’L ASS’N OF COUNTIES, supra note 143, at 9-20 tbl.II (summarizing

formsofcountygovernanceinallfiftystates).152. Seeid.at7.153. Seeid.154. Seeid.(describingthesevariations).155. Seeid.;seealso,e.g.,ARK.CODEANN.§14-14-801(a)(Repl.2013)(“[C]ounty

government, acting through its county quorum court,may exercise local legislative

executives.156 Many states allow counties to alter thestructureoftheirexecutiveauthorityorincreasethesizeofthegoverningbody.157

In addition to administrative responsibility, countygoverning bodies have legislative authority whereby theycan debate and adopt county ordinances.158 Some statesallow counties to pass ordinances on any issue notspecifically excluded fromcounty authorityby state law.159Other states, such as Colorado, give counties limited,enumeratedpowers.160

Theprocessforadoptingcountyordinancesgenerallyfollows the usual process for public lawmaking.161 Theprocess begins at a public meeting of the governing bodywhen a representative or citizen initiative introduces anordinance for consideration.162 Subject tomultiple-readingrequirementsinsomestates,thegoverningbodythenreadsand publicly debates the proposed ordinance.163 Theproposed ordinance may also be subject to committeereview or investigation.164 Significantly, all states permit(and some require) the governing body to allow for publiccomment regarding proposed ordinances.165 Once thegoverningbodyandthepublichavediscussedtheproposedordinance,membersofthegoverningbodypubliclyvoteon

authority not expressly prohibited by the Arkansas Constitution or by law for theaffairsofthecounty.”).

156. NAT’LASS’NOFCOUNTIES,supranote143,at7.157. Seeid.at7-8(discussingcountyhome-ruleauthority).158. Seegenerally id.at9-20(describinglegislativeauthorityofcountiesinall

fiftystates).159. See, e.g., ARK. CODE ANN. § 14-14-801(a) (“[C]ounty government, acting

through its county quorum court, may exercise local legislative authority notexpressly prohibited by the Arkansas Constitution or by law for the affairs of thecounty.”).

160. SeeCOLO.REV.STAT.ANN.§30-11-101(West2016).161. See, e.g.,TONYE.WINDHAM,UNIV.OFARK.DIV.OFAGRIC.,PROCEDURALGUIDE

FOR ARKANSAS COUNTY QUORUM COURT MEETINGS 12-14 (rev. ed. 2006) (describingArkansas’sprocedures).

162. Id.at12.163. Id.at13-14.164. Id.at17-18.165. See,e.g.,ARK.CODEANN.§14-17-207(b)(1)-(2)(Repl.2013).

the ordinance, and amajority vote is usually sufficient fortheordinancetopass.166

Significant for present purposes, all sixteen statesthatallowforconstitutionalamendmentbycitizeninitiativehavefunctioningcountygoverningbodieswithatleastsomelegislativeauthority.167

B. An Overview of the County-Ratification Model Stateswiththecitizeninitiativealreadyinplacecould

perhapsimprovethedemocraticqualityoftheiramendmentprocesses by requiring a majority (or a proportionallyweightedmajority)168ofthestate’scountygoverningbodiesto approve all citizen-initiated amendments rather thansubmittingthoseamendmentstoastatewidereferendum.

Tofocusmyanalysisandqualifymyarguments, it ishelpful to identifythekeyfeaturesofmyproposal. First, itwould not affect the legislative-referral method ofamendment. My suggestion is limited to only the citizen-initiative method in the sixteen states that allow foramendment by both legislative referral and citizeninitiative.169Thisisbecause,asothershavenoted,allowingamendments to be initiated by either the legislature ordirectlybycitizensprovidesauseful checkonabuseof theamendmentpower.170 Thereare,however,seriousflawsinthecitizen-initiativeprocessthatmustbeaddressedinorderto improve state amendment procedures. Thus, reform totheinitiativeprocessisthemostappropriateinquiry.

Second, my proposal does not offer any reforms tothesignature-gatheringphaseofthecitizeninitiative.Thisisnot to suggest that reform in thatarea isunwarranted,but

166. HENRYM.ROBERT,ROBERT’SRULESOFORDER339(rev.1970).167. SeeNAT’LASS’NOFCOUNTIES,supranote143,at28-64.168. As discussed in more detail below, the ratification threshold might be

adjusted to address Equal Protection concerns related to voter equality. See infraSectionV.A.

169. Seesupranote15(listingstatesincludedinmystudy).170. See Benjamin, supra note 84. The county-ratification model could be

applied to ratification of legislatively referred amendments, but I focus only on thecitizen-initiativeprocesstonarrowmyanalysis.

my focus here is on the ratification portion of the citizeninitiative because it has received very little scholarlyattention. Under my county-ratification model, citizenswouldinitiateamendmentsinthesamemannerascurrentlypermitted in initiative states. Inmost states, this requiresobtainingaspecifiednumberofsignaturesinsupportoftheproposal, and then submitting the text of the proposedamendment and proof of signatures to the state attorneygeneral for certification.171 After the attorney generalcertifies a citizen initiative, he or shewould then send theproposed amendment to all county governing bodies forconsideration.Theattorneygeneralwouldlikelyneedtosetadeadlineforcountiestoreturnan“up-or-down”responsetotheproposedamendment,buttheattorneygeneralwouldnot otherwise interfere with counties’ procedures fordecidingontheamendment.

Third, once the attorney general certifies citizeninitiatives,countygoverningbodieswouldreviewtheminamanner very similar to how they review ordinances. Theproposedamendmentwouldbeplacedon theagenda forapublicmeeting of the governing body. The bodymemberswoulddiscuss the amendment, solicit citizen feedback, andcouldevenrefertheamendmenttoacommitteeforfurtherinvestigation. After the governing body’s deliberations, itwould publically vote onwhether to approve or reject theamendment. Each county would report its collectivedecisionbacktotheattorneygeneral,whowoulddeterminewhether the required number of counties approved theamendment.

Obviously, this is a very general sketch of how acounty-ratification model might work. Deploying a modellikethisforanyparticularstatewouldrequireconsiderationof myriad other important design issues.172 However, itprovides a sufficient framework to begin an assessment of

171. See,e.g.,ARK.CODEANN.§7-9-107(Supp.2015).172. SeeinfraSectionV.E(discussing“implementationcomplexity”asacounter

argumenttothecounty-ratificationmodel).

how a county-ratification process might improve theinitiativeprocess.

C. How a County-Ratification Model Might Improve the Democratic Quality of State Amendment Processes

ThisSectionprovidesapreliminaryanalysisofhowacounty-ratification model might improve the democraticqualityofstateamendmentprocesses.Todothis,Idrawonexisting theoretical and empirical scholarship regardingdecentralization anddemocratic theory to explorehow themodel might affect amendment decision making in mysixteen-state sample. I advance three arguments in thisregard: (1) channeling amendment debates into countygoverning bodies might improve citizen participation inamendment politics; (2) channeling amendment debatesinto county governing bodiesmight improve the quality ofpublicdeliberation;and(3)dividingtheamendmentpowerbetweencountygoverningbodiesratherthanconsolidatingitinthelegislatureorastatewidereferendummightreducespecial-interest influence in amendment politics and moreeffectivelyempower“thepeople.”173

1. Improved Citizen Involvement in Amendment Decisions As noted above, citizens are increasingly alienated

fromthe initiativeprocessbywell-financedspecial interestgroups thatpay topdollar tosignature-gathering firmsandrun expensive statewide advertising campaigns.Additionally, many citizen voters learn about the proposalonly days before the vote, consult only one source ofinformationregardingtheinitiative,andveryrarelydiscussthe initiative with anyone.174 This represents a very thin,

173. My theoretical approach to citizen participation in amendment practice

generallyfollowsthemodeloflocalismandpoliticalparticipationsetoutbyGeraldE.Frug in The City as a Legal Concept, 93 HARV. L. REV. 1057, 1068-70 (1980). SeeRichardBriffault,OurLocalism:PartII—LocalismandLegalTheory,90COLUM.L.REV.346,393-95(1990)(describingFrug’sargumentregardingthedemocraticvirtuesoflocalism).

174. SeeGais&Benjamin,supranote1,at1301.

suboptimal level of political involvement by citizens. Thecounty-ratification model might help improve citizeninvolvement inamendmentpoliticsbyreducingdemocraticscaletoproviderealopportunitiesforcitizenengagement.

It is a truism in democratic theory that citizenparticipation tends to increase as jurisdiction sizedecreases.175Allthingsbeingequal,citizensaremorelikelyto engagewith small, local democratic institutions becausethe costs of participation are lower and the likelihood ofaffecting the outcome is greater. As a result, localdemocraciestendtobemoreresponsiveandaccountabletotheirconstituents’preferencesthanlargerdemocracies.176

Theempiricalresearchtestingthesepremisesisvastand somewhat contested.177 However, a few importantthemeshave surfaced. First, as jurisdiction size decreases,the primary effect on citizen participation seems to be anincrease in the quality of citizen participation.178 Forexample, voter turnout for local elections is notoriouslylow,179 but citizen involvement in “thick” forms of politicalparticipation—such as contacting officials, attendinghearings,andevenrunningforpublicoffice—isremarkablyhigh.180 In smaller jurisdictions, citizens are much morelikely to personally engage in these forms of politicalparticipation.181

175. See Roderick M. Hills, Jr., Federalism and Public Choice, in RESEARCH

HANDBOOK ON PUBLIC CHOICE AND PUBLIC LAW 207 (Daniel A. Farber & Anne JosephO’Connelleds.,2010);Frug,supranote173,at1069.ThisintuitionwasattheheartofJefferson’sWardRepublic.SeeMichaelHardt,JeffersonandDemocracy,59AM.Q.41,69-72(2007).

176. See J. ERIC OLIVER, LOCAL ELECTIONS AND THE POLITICS OF SMALL-SCALEDEMOCRACY16(2012)(“[G]overnmentalaccountabilityandaccessislowerinalargedemocracythanasmallone.”).

177. SeegenerallyDavidKarlsson,ADemocracyofScale,SizeandRepresentativeDemocracy in Swedish Local Government, 17 SCANDINAVIAN J. PUB. ADMIN. 7 (2013)(summarizingscaleliterature).

178. SeeHills,supranote175.179. See OLIVER, supra note 176, at 55 (stating voter turnout in most local

electionsisbelowtwenty-fivepercent).180. SeeHills,supranote175.181. SeeOLIVER,supranote176,at55-56;Hills,supranote175.

Identifying thepopulation “tippingpoint” for “thick”forms of citizen participation is complicated and involvesmanyvariables.182 Empiricistshavenevertheless identifiedsomebenchmarks. EricOliver’s 2001 seminal study foundthat thick citizen participation is highest in jurisdictions of5000or less, but also found that thick citizenparticipationremains meaningful in jurisdictions between 50,000 to250,000 and even in jurisdictions between 250,000 to onemillion.183 In local jurisdictions of more than one million,thickcitizenparticipationislesslikely.184

Oliver’s benchmarks are interestingwhen comparedtocountyandstatepopulationsacrossthesixteenstatesthathave the citizen initiative.185 Table 1 below shows thenumber of counties in each state that fall within Oliver’spopulation benchmarks, as well as the overall statepopulationforeachstate.186

182. SeeKarlsson,supranote177,at8-10.183. See OLIVER, supra note 95, at 42-45 (quantifying “nonelectoral civic

participation”injurisdictionsofvaryingsizes);seealsoROBERTA.DAHL&EDWARDR.TUFTE, SIZE ANDDEMOCRACY63-64 (1973) (discussing voterparticipation in Swedishcommunes).

184. SeeOLIVER,supranote95,at47.185. There are significant limitations on this comparison. Oliver primarily

studiedmunicipalities and other incorporated local governments that covered lessterritory and were perhaps more visible to citizens. Nevertheless, Oliver’sbenchmarks provide a helpful starting point in assessing the virtues of a county-ratificationmodel.

186. I compiled this table from the Census Bureau’s 2014 populationprojections, which are based on the 2010 Census data. See Estimates of ResidentPopulationChangeForCountiesandCountyRankings:July1,2013toJuly1,2014,U.S.CENSUS BUREAU, [hereinafter 2014 Population Projections], https://www.census.gov/popest/data/counties/totals/2014/CO-EST2014-03.html [https://perma.cc/B3PE-VCCT](lastvisitedMar.31,2016).

Table1

NumberofCountieswithPopulations

Between:

State StatePopulation

0-5,000

5,000-50,000

50,000-250,000

250,000-1mil.

Over1mil.

AZ 6,731,484 0 4 8 1 2AR 2,966,369 0 62 12 1 0CA 38,802,500 2 13 17 17 9CO 5,355,866 14 35 6 9 0FL 19,893,297 0 26 18 18 5MI 9,909,877 1 47 26 7 2MS 2,994,079 2 66 14 0 0MO 6,063,589 7 86 17 4 1MT 1,023,579 21 29 6 0 0NE 1,881,503 37 52 2 2 0NV 2,839,099 4 8 3 1 1ND 739,482 29 20 4 0 0OH 11,594,163 0 39 40 7 2OK 3,878,051 7 56 11 3 0OR 3,970,239 3 17 11 5 0SD 853,175 30 33 3 0 0

Lookingonly at jurisdiction size, countypopulations

across these states seem to fit nicely within the preferredsizesforthickcitizenparticipation.187Statepopulations,onthe other hand, seem much less likely to foster thickparticipation. Indeed, every state has a majority of itscountypopulationsunderonemillionpeople,butonlyNorthand South Dakota have a statewide population under onemillionpeople.188

187. Another factor that affects citizen participation is the significance of the

issuesatstake.Marshfield,supranote95,at1189.Thegreaterthesignificanceoftheissue,themorelikelycitizensaretogetinvolvedinthedecision-makingprocess.Id.This variable also suggests that a county-ratification model might facilitate bettercitizenparticipationbecause itwouldput constitutional issues in thehandsof localgoverningbodies.

188. See 2014 Population Projections, supra note 186 (follow “NorthDakota”hyperlink);id.(follow“SouthDakota”hyperlink).

Counties therefore seem to be preferablejurisdictions within which to foster more direct andmeaningful citizen participation. If amendment decisionswereputon theagenda forresolutionbycountygoverningbodies,itseemsthatcitizenswouldbemorewillingandableto engage the process in meaningful ways, such as byattending hearings, contacting local representatives, andevenrunningforcountyofficethemselves.Alloftheseformsof participation become realistic (and even more likely)when amendment decisions are moved from a statewidereferendumtolocaljurisdictions.

There are, of course, some extraordinarily largecounties. LosAngelesCounty, forexample, is the largest inthe country with more than ten million residents,comprisingapproximatelytwenty-sixpercentofCalifornia’soverallpopulation.189 SendingamendmentdecisionstoLosAngeles County might not enhance citizen participationwithin the county because of its large size. However,California has thirty-two other counties with populationsbelow 250,000 people.190 Citizens in those counties arelikely to engage with amendment decisions in moremeaningfulwaysthaninastatewidereferendum,whichmaymakeacounty-ratificationmodelvaluablenonetheless.

2. Improved Public Deliberation Regarding Amendment Decisions

In addition to “thickening” citizen participation, acounty-ratification model might also strengthen thedeliberative quality of amendment decisions. As notedabove,oneof themajorproblemswiththecitizen initiativeis that it circumvents meaningful deliberation regardingconstitutional changes, and can even enable collectiveprejudicesby allowing amendmentdecisions tobedecidedinthesecrecyoftheballotbooth.

189. Seeid.(follow“California”hyperlink).190. Id.

These problems might be mitigated by requiringcounty governing bodies to debate and transparently voteon proposed amendments. The county-ratification modelenvisions a county governing body voting on proposedamendmentspursuant to similarprocessesandproceduresforadoptingcountyordinancesinmanystates.Theprocessusually includes review by body members, reasonableopportunitiesforcommitteeinvestigationsandreports,anda public hearing. The process would culminate with atransparent vote by locally elected representatives onwhether to ratify or reject proposed amendments. Thisprocess isdramaticallydifferent thanthecurrent initiative-ratification process because it would subject amendmentdecisions to public review and discussion by locallyaccountable representatives. There are at least two likelybenefitstothismoredeliberateandtransparentratificationprocess.

First, it can help mitigate majoritarian abuses.Allowing amendments to be ratified by aggregating secretindividual preferences seems to invite expressions ofprejudice. However, when discriminatory viewpoints aresubject to public exposure and discussion, they are morelikely to be moderated, vetted, and even rejected.Representative decision making is especially effective atmitigating majoritarian abuses when minority viewpointshave at least some voice in the process. To be sure,discrimination can occur even in representative decision-making bodies—especially in homogenous bodies whereminorityviewpointsarenotexpressedortakenseriously—but representative decision making seems a vastimprovement on the aggregation of secret privatepreferences.

Regarding minority viewpoint representation,counties have the promise of being effective politicalsubdivisions for giving voice tominority viewpoints in theamendmentprocess.Usingraceasaproxyfordiversity,thechartbelowshowsthenumberofcountieswithpopulationsthataremoreraciallydiversethanthestate’spopulationasa

whole as well as the number of majority-minoritycounties.191

As thesedatashow,everystatehasseveral counties

with greater minority representation than in the statepopulation as a whole, and every state except Ohio andOregon have at least onemajority-minority county. This issignificant for deliberative purposes because itmeans thatunlike a statewide referendumwhere minority viewpointsarelikelywashedoutbysimpleaggregationofprivatevotes,acounty-ratificationmodelwouldgiveminorityviewpointsa formal, public structure for voicing opinions regarding

191. IcompiledthischartfromCensusBureauDataonracialdiversityfromthe2010 census. In classifying minority and majority populations, I followed thedefinitionsandclassificationsusedbytheCensusBureau. SeeLINDSAYHIXSONETAL.,U.S. CENSUS BUREAU, THE WHITE POPULATION: 2010, at 16-17 (2011),http://www.census.gov/prod/cen2010/briefs/c2010br-05.pdf [https://perma.cc/DWB4-UYRT] (theminoritypopulation includes “peoplewho reported their raceandethnicityassomethingotherthannon-HispanicWhitealone”).

10152025303540

MinorityCoun;esinIni;a;veStates

NumberofCoun-esMoreDiversethanStatewidePopula-on

NumberofMajority-MinorityCoun-es

proposed amendments.192 Minority representatives oncountygoverningbodieswillhaveanopportunitytoexpressminority viewpoints during the body’s deliberations.Additionally,governingbodiesinmajority-minoritycountieswill have an opportunity to publically decide on proposedamendments from a minority viewpoint and providejustificationsfortheirdecisionsthatothergoverningbodiesmay then consider. A transparent public forum that givesvoice tominority communities can facilitate accountabilitybetween counties, promote inclusivedeliberation, andhelpmitigatemajoritarianabuses.

In addition to mitigating majoritarian abuses, acounty-ratification model might improve the accuracy anddissemination of information regarding an amendmentproposal. One of the primary benefits of representativedecision making is that it provides a public forum for theexchange and vetting of viewpoints on a proposal. Ascitizens and representatives discuss a proposal, collectiveknowledge regarding the proposal presumably grows andthe overall quality of the decision-making processconsequently improves. A county-ratification model canfacilitate this process by giving citizens a forum at publichearings to express viewpoints on proposed amendmentsandbyforcingrepresentativestodiscussamendmentswitheachotherandthepublic.193

In sum, ratifying amendmentsbypublic referendumseems to result inunder-inclusiveand limiteddeliberation.A county-ratification model might improve the process by

192. Empiricistshavefoundthatdiversitytendstodecreaseasjurisdictionsize

decreases.SeeOLIVER,supranote176,at17(chartingracialdiversitybutnotingthat“[t]hesametrendoccurswithnearlyeveryothersocialcategory”).

193. County governments appear to have a strong tradition of gatheringinformationregardingpublicopinionbeforemakingdecisions. SeeMaureenBerner,Citizen Participation in Local Government Budgeting, POPULARGOV’T, Spring 2001, at23-24 (discussingNorth Carolina public budget hearings). A 2001 study of citizenparticipationincountybudgingapprovalfoundthatcountiesuseavarietyofdifferentmethods to solicit citizen feedback, such as informal “[c]offeehouse conversations”,“[v]isitstolocalcivicgroups”,andtheformationof“[c]itizenadvisoryboards.”Id.at27tbl.6.

providing a voice for minority viewpoints and facilitatingmorerobustdiscussionregardingproposedamendments.

3. Reduced Special Interest Influence in Amendment Decisions

As noted above, amajor problemwith the initiativeprocessistheinfluenceofwell-financedspecialinterests.194Thesegroupshaveunderminedtheinitiativeasagrassroots,populist institution. Instead, the initiative often involvesreforms with limited popular support, but with significantfinancial backing from corporate or ideological interestgroups.

Acounty-ratificationmodelmighthelpreducespecialinterest influence and restore the initiative to its popularsovereignty roots. From a democratic theory perspective,this might seem counter-intuitive. The traditionalMadisonian assumption is that lower levels of governmentaremoresusceptibletocapturebyspecialinterestsbecauseof, among other things, greater cohesiveness of interestgroupsandsmallerpopulationsizes.195

However, this assumption has come under criticismby recent theoretical investigation.196 The susceptibility tocapture of low-level governing institutions depends on avariety of variables such as the relative extent of politicalcompetitionanddifferencesinelectoralsystems.197Itisnotclear, therefore, that county governments are alwaysmoreprone to capture than state institutions.198 The “relative

194. Hoesly,supranote2,at1202-03.195. THE FEDERALISTNO. 10, supranote 100; see also PranabBardhan&Dilip

Mookherjee,CaptureandGovernanceatLocalandNationalLevels,90AM.ECON.REV.135,135,139(2000)(describingtheMadisonianview).

196. See Bardhan & Mookherjee, supra note 195, at 135 (explaining that“[d]espitetheimportanceofthisissue,notmuchsystematicresearchappearstohavebeen devoted to assessing the relative susceptibility of national and localgovernmentstointerest-groupcapture”).

197. Id. at 139 (“[T]he relative proneness to capture of local governmentsdependsonamultitudeofdiversefactors.”).

198. Id.

capture at the local level may... be context- and system-specific.”199

More importantly, however, the county-ratificationmodel would not centralize the amendment power in thehandsofoneparticularcountygoverningbodythatcouldbeeasily captured. Instead, it would divide the amendmentpower between the various counties within a state. Myhypothesis is that this design arguably involves a moreeffective check on capture and abuse than centralizing theamendmentpowerinastatewidereferendum.200Itisbasedon Madison’s checks-and-balances argument in FederalistNo. 51, where Madison argued that the tyranny of themajoritycanbemanagedbydividingsociety “intosomanyparts, interests, and classes of citizens, that the rights ofindividuals, orof theminority,will be in littledanger frominterestedcombinationsofthemajority.”201

Statewide referenda represent an absoluteconsolidation of the amendment power with a majority ofvoterswhowill decide the issue in secret andwithout anydeliberation. Referenda are therefore relatively easy tocapture through mechanisms such as mass media.202 Acounty-ratification model, on the other hand, divides theamendment power between myriad multi-member countygoverning bodies. This division and separation of powerwithin the amendment process could help protect againstcaptureandmajoritarianabusesinseveralways.

First,itwouldseemmuchmoredifficulttocaptureamajority of the county governing bodies in a state than toaggregate public opinion.203 Voter decision making on

199. Id.200. SeeJonathanL.Marshfield,DecentralizingtheAmendmentPower,19LEWIS

&CLARKL.REV.963(2016)(makingthesamesuggestionregardingdecentralizationofnationalamendmentprocessestoincludesubnationalunitssuchasstates).

201. THEFEDERALISTNO.51,at121(JamesMadison)(MichaelA.Genoveseed.,2009).

202. SeeOLIVER,supranote176,at20-21(discussingtheroleofmediainlargejurisdictionalpolitics).

203. See S. CandiceHoke,PreemptionPathologies andCivicRepublicanValues,71B.U.L.REV.685,693(1991)(“OnemaypositthatitismoredifficulttocapturefiftystatelegislaturesandbureaucraciesthantomasteroneinWashington.”).

statewide referenda is generally simplistic and ideological,which means that interest groups can impact referendaoutcomes by running mass media campaigns that play onsimplistic ideological associations. Decision making andaccountability regarding local governing bodies areconsiderably more personal, deliberative, andtransparent.204 Under a county-ratification model, localrepresentatives would have a personal interest inamendment decisions because they would be openlyaccountable to constituents for theirvote.205 This suggeststhat they would have greater personal incentives to studyanddiscussproposedamendmentsthancitizensvotinginareferenda, which could create a check on capture and anincentiveforreducingagencycosts.

Second, a county-ratification model would create acompetitive environment between counties, which couldhelp limit capture. By giving each county a role in theamendmentprocess, countieshavean incentive tomonitoreach other. Capturing the process would requireovercoming a complicated coordination problem. One“captured” county would likely be exposed by the othercounties as acting under the disproportionate influence ofspecialinterests.Tocapturetheprocessasawhole,specialinterests would therefore have to obtain a degree ofcoordination between counties. This would likely makecapturemoredifficultthaninastatewidereferenda.

Third,acounty-ratificationmodelhasthepromiseofrestoringadegreeofpopularsovereigntytotheamendmentprocess. Although the referenda was intended as amechanism for restoring popular sovereignty,206 itssusceptibility to capture has significantly undermined itspopular legitimacy. Local democratic institutions, on theotherhand, aregenerally considered tobevery responsive

204. SeeOLIVER,supranote176,at20-21(discussinghowpoliticalactivismand

accountabilityinsmallerjurisdictionsisoftenlessmediated,morepersonal,andlessideological).

205. Seeid.206. Magleby,supranote80,at13-15.

and accountable to constituents.207 This is becauserepresentatives are “closer” to constituents in smalldemocracies, which means that political life tends to bemore interpersonal and lessmediated throughdevices likemedia and political parties.208 To the extent countyrepresentativesaremoreaccessibleandresponsiveto localcommunities in any given state, delegating the amendmentauthority to county governing bodiesmight be an effectivemeans of taking the amendment power back from specialintereststhatcontrolthereferendumandrestoringittothepeople.

V. PROBLEMS WITH INCLUDING COUNTY LEGISLATIVE BODIES IN STATE AMENDMENT

PROCESSES The county-ratification model will not solve all

problemsrelatedtostateconstitutionalamendmentprocess.In fact, it surely comes with its own costs. In this Part, Iidentify some of the major problems and likely counter-argumentstoacounty-ratificationmodel.

A. Equal Protection Asignificantissuewithacounty-ratificationmodelis

compliance with the one-person, one-vote principle underthe Fourteenth Amendment’s Equal Protection Clause.209The Fourteenth Amendment requires equality in thepopulationoflegislativedistrictssothatanindividual’svotein one district does not count substantially less than anindividual’s vote inanotherdistrict.210 TheSupremeCourthas applied this principle to state and local elections, but

207. See OLIVER, supra note 176, at 17-18; Karlsson, supra note 177, at 8-9;

RusselM.Lazega&CharlesR.Fletcher,ThePoliticsofMunicipalIncorporationinSouthFlorida,12J.LANDUSE&ENVTL.L.215,225(1997).

208. SeeOLIVER,supranote176,at17-18.209. SeeBakerv.Carr,369U.S.186,187-88(1962);REYNOLDS,supranote143,

at67-75(discussingone-person,one-voteinrelationtolocalgovernment).210. Reynoldsv.Sims,377U.S.533,568(1964).

never directly to state constitutional amendmentprocedures.211

The few lower courts that have addressed the issuehave held that there is “no rational basis to distinguishbetween voting on representatives in the legislature, andvoting on constitutional amendments.”212 Thus, althoughnotdefinitivelydecidedbytheSupremeCourt,thedominantposition is that amendment ratification must comply withtheone-person,one-voterequirement.213

Assumingthattheone-person,one-voterequirementwould apply to the county-ratificationmodel,214 themodelraises someequality concerns thatwould likelyneed tobeaddressed. Although county representatives must alreadybe elected consistent with the one-person, one-voterequirement within each county,215 the county-ratificationmodelwoulddecideastatewideproposalbyaddingup thenumber of counties that approve the proposal. If there isgreat disparity in populations between counties (which isusually true), there could be significant under-representation for voters in populous counties and over-representationforvotersindesolatecounties.

Arizona provides a helpful example. Arizona isdivided into fifteen counties, but sixty-one percent of theoverallpopulationlivesinonecounty(MaricopaCounty).216Thesmallestcounty,GreenleeCounty,containsanestimated

211. See Peter Suber, Population Changes and Constitutional Amendments:

VersusDemocracy,20U.MICH.J.L.REFORM409,420n.19(1987).212. See,e.g.,Stateexrel.Wittv.StateCanvassingBd.,437P.2d143,150,155

(N.M.1968)(holdingthatNewMexico’samendmentrulesviolatedEqualProtectionbecause they required amendments to be ratified by majorities in a majority ofcounties).

213. SeeWestv.Carr,370S.W.2d469(Tenn.1963)(discussingtheone-person,one-voteprincipleinthecontextofaconstitutionalconventionprocess).

214. There are perhaps arguments that conventional Equal Protectionprinciples should not apply to a county-ratification model, but I leave thoseargumentsforanothertime.

215. Reynolds,377U.S.at568.216. AnnualEstimates of theResidentPopulation for Counties:April 1, 2010 to

July 1, 2015, U.S. CENSUS BUREAU, https://www.census.gov/popest/data/counties/totals/2015/CO-EST2015-01.html [https://perma.cc/BZ74-YPVC] (follow“Arizona”hyperlink).

9529 people, or approximately 0.14% of the overall statepopulation.217Clearly,givingeachcountyanequalsayintheratificationprocessunderthosecircumstanceswouldresultinsignificantinequalitiesforvoters.

Theequalityproblemcouldpossiblybeaddressedbygiving each county a weighted vote in the ratificationdecisionthatisproportionaltoitsshareoftheoverallstatepopulation.218 However, this could affect the dynamics ofthemodel.InArizona,forexample,aweighted-majorityrulethat required only a simple majority of the population toapprove an amendment would mean that the governingbodyforMaricopaCountywouldcontroltheoutcomeforthewhole state.219 To ensure that all counties have somerepresentation in theprocess,Arizonawould likelyneedtoadopta super-majority requirement thatwouldnecessarilyincludeothercounties intheprocessandpreservemanyoftheproceduralvirtuesofthecounty-ratificationmodel.220

ItisnotpossibletofullyanalyzetheEqualProtectionissue here. This is a complex issue that deserves moreattention if the county-ratificationmodel is considered forany particular state. My purpose is only to raise the issueand itspotentialeffectson the lawfuldesignof thecounty-ratificationmodel.

B. Mismatched Expertise and Limited Resources An important objection to the county-ratification

model is that county representatives arenotwell suited todeciding important statewide issues. Countyrepresentatives are usually elected based on theirmanagerial capacity or performance, and not theirideologicalviewpointsorstatesmanship.221 This isbecause

217. Id.(providingestimated2015Arizonacountypopulations);id.(follow“AllStates”hyperlink)(providingtotalestimated2015statepopulations).

218. Anothersolutionmightbetomakecountyratification“advisory,”followedbyabindingpublicreferendum.ThankstoJustinLongforsuggestingthis.

219. Seesupratextaccompanyingnote216.220. The Supreme Court has upheld super-majority requirements. See, e.g.,

Gordonv.Lance,403U.S.1,7-8(1971).221. OLIVER,supranote176,at6-7.

the primary business of county government relates to theprovision of basic government services rather than theresolution of ideological conflicts.222 County governingbodies also lack the breadth of perspective on stategovernment that a state legislaturemight possess.223 Thismight make county governing bodies a poor forum fordebating constitutional amendments,which often touch onbroad ideological issues that have myriad compleximplicationsforastateanditsgovernment.224Additionally,county governments are notoriously under-resourced.225Adding controversial amendment decisions to the countyagenda could further exhaust county government anddetractfromtheeffectiveprovisionoflocalservices.

This critique raises important considerationsregardingthecounty-ratificationmodel.However,itiseasyto overstate these concerns and lose sight of the currentstateofamendmentpracticeininitiativestates.Thecounty-ratification model is not a perfect system for decidingamendment issues. There are surely limitations and costsfor enlisting county governing bodies in the amendmentprocess, but those costsmust be compared to the existingprocessandweighedagainstpossibleimprovementstothatprocess.

For example, although county representativesmightbe less qualified to debate many amendment issues thanstaterepresentatives, theyareperhapsbettersituatedthancitizens voting in private, based on thin sources ofinformation. County representatives are increasinglysophisticated administrators because of the increasing roleof county government.226 Counties manage large budgets,adopt and implement complicated economic development

222. Id.223. Seeid.224. See id. at 10 (stating that local government is not an ideal venue for

resolvingdeepsocialandeconomicissues).225. SeeSalant,supranote146,at9;Lobao&Kraybill,supranote147,at253-

54.226. SeeLobao&Kraybill,supranote147(explainingthatcountyofficialshave

significantresponsibilities).

policies, administer federal and state assistance programs,coordinate regional planning, and act as intermediariesbetween citizens, municipalities, state government, andfederal agencies.227 Thus, county representatives arebecoming increasingly skilled at public administration andlikely have the ability to engage in a constructive publicdebateregardingmanyamendmentissues.

The point regarding county resourcesmight also beoverstated. Over the last decade, initiative states haveaveraged about 1.5 proposed amendments by initiativeevery two years.228 If California is removed from thatsample,initiativestateshaveaveragedonlyoneamendmentevery two years.229 Thismeans that, on average, countieswouldhavehadtoconsideronlyoneamendmenteverytwoyears.230 Although a county-ratification model will surelydetract some county resources from other activities, thisdoesnotseemlikeasignificantimpositiononcounties.

C. The People Rarely Give Up the Initiative Afurthercriticismofthecounty-ratificationmodelis

that it is the equivalent of a constitutional “pipe dream”becausestatesveryrarelydisposeoftheinitiativedespiteatraditionofredesigningamendmentprocedures.231

Itistruethatitispoliticallyverydifficultforstatestoeliminate the initiative. Once citizens have the right toamendtheirconstitutionsdirectly,theyareveryreluctanttogive that up.232 However, as John Dinan has noted, thetwenty-first century is characterized by a trend inamendmentdesigntowardslimitingtheinitiativeprocess.233

227. Seeid.228. Seesupranote75andaccompanyingchart.229. Seeid.230. Seeid.231. See Elizabeth Garrett & Mathew D. McCubbins, The Dual Path Initiative

Framework, 80 S. CAL. L. REV. 299, 310 (2007) (“Although one response to theproblems in the initiative process is to eliminate direct democracy, we do notadvocatethatpath....[W]edonotbelieveitisrealistic.”).

232. Id.233. JohnDinan,Twenty-FirstCenturyDebatesandDevelopmentsRegardingthe

DesignofStateAmendmentProcesses,69ARK.L.REV.283,292(2016).

Thus, it is possible that “tides” are changing and stateconstitutional designers would be more receptive to anamendmentprocessthatsignificantlyreformstheinitiative.

Additionally, the county-ratification model has thebenefit of retaining a grassroots form of constitutionalratification.Thus,votersmightbemoreinclinedtoconsiderit even if they would reject proposals to categoricallyeliminate the initiative. In light of the real problems incurrent initiative practice, the county-ratification modeldeservesconsiderationanddiscussion.

D. Making the Constitution Too Hard to Amend Anotherobjectiontothecounty-ratificationmodel is

thatitwouldmakestateconstitutionspracticallyimpossibleto amend through the initiative process. Becauseamendments would be subject to debate and approval incountygoverningbodiesacrossthestate,thereisreasontobelieve that amendment processes might become veryprotracted and difficult to navigate. Practically, thismightmean that state constitutions would become too stagnantand outdated or that the legislature’s power in theamendmentprocesswouldgrowtoogreatandunchecked.

This criticism raises important concerns, but therearecountervailingconsiderations.First,itassumesthattheconstitutional flexibility provided by current initiativepracticeisnormativelydesirable.Thereare,however,manywho argue that the special-interest-driven flexibility of theinitiativeprocess isdysfunctional,andthatgreaterstabilityisrequired.Inthatcase,thecounty-ratificationmodelmightbedesirabletoamendmentreformers.234

It is also hard to predict what effect the county-ratificationmodelwillhaveonconstitutionalflexibility.Themodel would likely make amendment by initiative moredifficultthanareferendummodel,butthereisnoreasontobelievethatitwouldmaketheinitiativeunusable.Irecently

234. Again, this seems to be the trend in contemporary debates by

constitutionaldesignersabouthowtoreformtheinitiativeprocess.Seeid.

concluded a study of decentralization in nationalconstitutional amendment rules and found countries thatrequire subnational units to participate in amendmentprocessesoftendonotexperienceloweroverallamendmentrates.235

Constitutional flexibility is a very importantconsideration in designing amendment rules. A county-ratification model would likely make amendment byinitiativemore difficult than a referendummodel, but thismaybeadesirableoutcomeoranacceptablecost forsomestatesinreformingtheiramendmentprocedures.

E. Implementation Complexity Finally, the county-ratification model would be

relativelycomplextoimplementinanygivenstate.Itwouldrequireconstitutionalamendmentsandprobablylegislationonavarietyofancillaryissues. Constitutionalamendmentsmight have to change provisions besides the amendmentrules. ThelegislativearticleunderCalifornia’sconstitution,for example, would likely have to be amended because itstatesthat“thepeoplereservetothemselvesthepowersofinitiative and referendum.”236 This sort of constitutional“housecleaning”canbecomplexwhensignificantstructuralchangesaremade.

Additionally,manystateswouldhavetochangelocalgovernment legislation to accommodate the county-ratification model. States that allow county executives toveto county ordinances and resolutions might want toeliminate that veto authority for amendment decisions.237Therearelikelymanyotherlegalcomplexities.InArkansas,for example, the legislature would likely need to considerwhether to change the County Government Code so that

235. SeeMarshfield,supranote200.236. SeeCAL.CONST.art.IV,§1.237. See WINDHAM, supra note 161, at 15-16 (discussing executive veto

authorityinArkansascountygovernance).

amendment decisions are not subject to county referendalikeordinarycountyordinances.238

Thereisnodenyingthatthelegalaspectsofadoptingthecounty-ratificationmodelarecomplex. Anyjurisdictionconsidering the model would have to weigh whether theimplementation complexity outweighs the other likelybenefits. Mypurposehere isonlytoacknowledgethe legalcomplexityassociatedwithmyproposal.

VI. CONCLUSION State constitutional amendments contribute greatly

to constitutional life in theUnited States. As of January 1,2015, voters have considered 11,369 amendments tocurrently operative state constitutions.239 Voters ratified7481ofthoseamendments.240Thisrepresentsasignificantcorpus of positive constitutional law created through stateamendment processes. It is important, therefore, thatscholars give more attention to understanding how theseprocessesarefunctioningandcanbeimproved. Mygoal inthis essay is to encourage creative thinking regarding thedesignofstateamendmentprocesses,andalsosuggestthata county-ratification model of some kind might be animprovementonthecitizen-initiativeprocess.

238. See WINDHAM, supra note 161, at 13 (describing the rules for local

referendainArkansas).239. See Dinan, supra note 2, at 11 tbl.1.1. Delaware’s 145 amendments are

includedinthisnumbereventhoughtheywerenotsubmittedtothevotersdirectly.240. Seeid.

IMPROVING AMENDMENT

Documents