schedler 1999 conceptualizing accountability

8/12/2019 Schedler 1999 Conceptualizing Accountability

1/12

heSELF RESTRAINING STATE

Power and Accountabil ityin New Democracies

edited b yAndreas Schedler

Larry DiamondMarc F Plattner

L~ERI EN N ERP U B L I S H E R S

1 \ U L D ERI N o N


2/12

pub li sh ed in th e Uni ted S ta te s o f Amer ic a in 1999 byLynne Rienner Publishers, Inc.1800 30th Street, Boulder, Colorado 80301and in the Uni ted Kingdom byLynne Rienner Publishers, Inc.3 Henrietta Street, Covent Garden, London WC2E 8LU 1999 by Lynne Rienner Publishers, Inc . andThe Nationa l Endowment fOI Democracy . All r ights rese rved by the publisher.

Library of Congress Cataloging-in-Publication DataThe se lf -r es tr aining s ta te : p ower a nd ac coun ta bi li ty in n ewdemocracies / edited by Andreas Schedler, Larry Diamond, and Marc EPlattner.p. cm.Includes bibliographical references and index.ISBN 1-55587-773-7 (he : alk. paper)ISBN 1-55587-774-5 (pbk : alk. paper)

l. Political corruption-Prevention. 2. Responsibility.3 . S ep arat io n o f powers. 4. Democracy. 1. Schedler, Andreas,1964- . ll. Diamond, Larr y Jay. I ll . Pl at tner, Mar c E, 1945-JF229.S47 1999352.3'5-DC21 98-44668CIP

British Cataloguing in Publication DataA Cataloguing in Publication record for t his bookis available from the British Library.Printed and b und in th nlt d 101 s o r Am rI ti

ontents

Aeknowledgments I X

1 IntroductionLarry Diamond, Mare F Plattner, and Andreas Sehedler 1

Part I Conceptual and Normative Issues2 Conceptualizing Accountability

Andreas Sehedler 133 Horizontal Accountability in New Democracies

Guillermo ODonnellComments on O'Donnell Democracy and Constitutionalism

Riehard L. Sklar The. ~imits of Horizontal AccountabilityPhilippe C . Sehmitter Tradition of AccountabilityMar F. Plattner A R sp ns t M y mrnentat rs

;I/II/ III/) 1 onnell 8

94

535963

I \


3/12

vi Contents

Part Electoral Administration5 A Brief History of Electoral Commissions

Robert A. Pastor6 Off the St ree ts and into the Courtrooms:Resolving Postelectoral Conflicts in Mexico

Todd A. Eisenstadt7 Institutionalizing Credible Elections in GhanaE. Gyimah-Boadi8 The Third Dimension of Accountability:The International Community in National Elections

Robert A. Pastor

Part 3 Judicial Systems9 A Brief History of Judicial Review

Herman Schwart:10 Judicial Independence and Judicial Reform in Latin AmericaPilar Domingo

11 Bui lding Judicia l Independence in Common Law AfricaJennifer Widner

12 Surpri sing Success:The New Eastern European Constitutional CourtsHerman Schwart:

Part 4 Corruption Control13 A Brief History of Anticorruption Agenc ies

Michael Johnston14 Corrupt ion, Democracy, and Reform in Benin

John R. Heilbrunn15 Combating Corruption in South Korea and Thailand

Jon S. T. Quah16 The Global Coalition Against orruption :

valuatin Transpar n y lnt rnutionn lFredr ik f/, l U /7 ,H and ./ / 11I f > / 1 1

75

83105

123

145151177

195

217227245

7

Contents vii

Part S Central Banks17 A Brief History of Cen tra l Bank Independencein Developing Countries

Sylvia Maxfield 28518 MisguidedAutonomy: Cen tral Bank Independence

in the Russian TransitionJuliet Johnson 29319 Learning from Failure: The International Financial Institutionsas Agencies of Restraint in Africa

Paul Collier 313

Part 6 Conclusion20 Restraining the State: Conflicts and Agents of Accountability

Andreas Schedler 333

BibliographyThe ContributorsIndexAbout the Book

351381385395


4/12

2Conceptualizing Accountability

ANDREAS SCHEDLER

W hat is the essence of politics? What is the key variable of politicalsc ience? Common sense gives us a c1ear answer: it is power. But asthe ear lier c1ass ical theoris ts knew: in polit ics, f irst comes power, then theneed to control it. In framing a government ... the great difficulty lies inthis: you must first enable the governmen t to cont rol the governed; and inthe next place oblige it to control itself. l Beginning with the ancientphilosophers, political thinkers have worried about how to keep powerunder control, how to domesticate it, how to prevent its abuse, how to sub-ject it to certain procedures and rules of conduc t.s Today, it is the fashion-able term accountability tha t expresses the cont inuing concern fo r checksand overs ight , for surveil lance and ins ti tutional constraints on the exerciseof power. All over the world (wherever the term is halfway translatable),international financial institutions, party leaders , grassroots activists, jour-nalists, and poli ti cal scientists have discovered the blessings and adhere tothe cause of public accountability. 3Without doubt , the term sounds appeal ing. Its fie ld of appl icat ion i s asbroad as i ts potential for consensus. And its semantic root, the notion ofaccounting, is nicely ambivalent ; i t evokes narrative accounts as well asbookkeeping. But do we know what it means? Are we c1ear about itssemantic boundaries and do we comprehend its interna l structu re? Not sur-prisingly, my answer is no: due to its relative novelty, accountability repre-sents an underexplored concept whose meaning remains evasive, whoseboundaries are fuzzy, and whose internal structure is confusing.This chapter does not set out to relate the concept of accountability tothe history of thought. Chapter 3 by Guillermo O'Donnell and the related

mm nts by Ri hard klar, Philipp Schrnitt er, and Marc Plattner accom-pllsh t\ cl d al of 8U h phi losophi al r fl cti n, Th pr et nse of this


5/12

nswerabilityAs with most terms we use in everyday language, we usually assume thatwe understand what we say when we talk about accountability, and thatothers do so as well . Related discuss ions therefore regular ly proceed on thebasis of implicit understandings, without recourse to any explicit definitionof the concept. Yet whenever authors do define the term explicitly, theytend to associa te i t wi th answerability as it s closest synonym.? Account-abil ity, one encyclopedic definit ion tell s us, i s the abil ity to ensure thatoffic ial s in government are answerable for their act ions.> We think thisprovides a va lid starting point. The not ion of answerabil ity indicates tha tbeing accountable to somebody implies the obligation to respond to nastyquest ions and, vice versa, that holding somebody accountable implies theopportunity to ask uncomfortable question .But what kind of answer t what kind f questi ns? In principIe,a untin a nci s may ask H unta] I \ t J S r J tw kinds f thin s.111 y nlfly ith J ask lh m t Ilvol/11 llholll lh s lr d ~ions o, Ih y muy askl h 11 r o I.~II/I/I th ir ' t i 1 1011 , l h 1111 I 11 th r o 1 1 It 1 r 1'(. t h

ConceptualizingAccountability 5 Conceptua l and Normative Issues

The attract iveness of accountabil ity der ives from its breadth. The concept isa comprehensive one. Rather than denoting one specific technique ofdomest ica ting power, it embraces three di fferent ways of prevent ing andredressing the abuse of poli tical power. It implies subjecting power to thethreat of sanctions; obliging it to be exercised in transparent ways; andforcing it to justify its acts. 1 treat the first dimension under the headingenforcement and sum up the last two dimensions under answerability.The three aspects together-enforcement, monitoring, and justification-turn polit ical accountabil ity into a mult ifaceted enterpr ise that copes with aconsiderable var iety of actual and potential abuses of power.

informationa l dimension of accountabi li ty) or fo r val id reasons (the argu-mentative dimension of accountabil ity) . Accountabili ty thus involves ther ight to receive information and the corresponding obligation to release allnecessary deta ils. Bu t it a lso impl ies the right to rece ive an explanation andthe corresponding duty to justify one's conduct.eOn the one s ide, exercising accountabil ity therefore involves elementsof monitoring and oversight. I ts miss ion includes f inding facts and generat-ing evidence. On the other side , the norm of accountabi li ty continues theEnlightment's project of subjecting power not only to the rule of law butalso to the rule of reason. Power should be bound by legal constraints butalso by the logic of public reasoning. Accountability is antithetical tomonologic power. I t establishes a dialogic relat ionship between account-able and accounting actors. It makes both parti es speak and engages themboth in public debate. It is therefore opposed not only to mute power butalso to unilateral speechless controls of power. 7The etymological ambivalence of accountabil ity, which s tems from thedouble connotation of accounts, sustains both dimensions. Financialaccounts contain detai led information prepared according to cer tain s tan-dards of classi fica tion and accuracy that make them both readable and re li-able for outsiders. Narrative accounts are legitimating tales that presentsome relieving end as the compelling outcome of previous developments.Or more formally, they represent definitional transformations of contestedaction, explanations that excuse or justify questionable behavior byproposing a normative s tatus for the behavior. 8 The informational dimen-s ion of accountabil ity relates to the f irst connotation of accounts-book-keeping; its discursive or explanatory dimension relates to the secondone-s-storytelling.?

chapte r i s more l imited. It aims at reconstruct ing the meaning of the con-cept as we currently use it. In essence, i t claims that the notion of polit icalaccountability carries two basic connotations: answerability, the obligationof publ ic offi cia ls to inform about and to explain what they are doing; andenforcement, the capac ity of accounting agenc ies to impose sanct ions onpowerholders who have violated their publ ic duties. This two-dimensionalst ructu re of meaning makes the concept a broad and inclusive one that,within its wide and loose boundaries, embraces (or at least overlaps wi th)lots of other terms-surveillance, monitoring, oversight, control, checks,restraint, public exposure, punishment-that we may employ to describeeffor ts to ensure that the exercise of power is a rule-guided enterpr ise.

A Two Dimensional Concept

nforcementUp to now we have described the exercise of accountabil ity essential ly as adiscursive activity, as a sort of benign inquiry, a friendly dialogue betweenaccounting and accountable parties. Yet answerability, and the double questfor information and justification it implies, is not the whole story ofaccountability. Political accountability involves more than the generation ofdata and the interplay of arguments. In addit ion to its informational dimen-sion (asking what has been done or will be done) and its explanatoryaspects (giving reasons and forrning judgments) , i t also contains elementsof enforcement (rewarding good and punishing bad behavior). It impliesthe idea that accounting actors do not just call into question but alsoeventually punish improper behavior and, accordingly, that accountablepersons not only tell what they have done and why, but bear the conse-qu nc f r it, including eventual negative sanctions.t? Exercises ofa ounruhilhy (hat xpos misd ds but do not impose material conse-


6/12

16 Conceptua l and Normative Issues

We may sum up the preceding reflections in the following one-sentencedefinition of accountability: A is accountable to B when A is obliged toinform B about A's (past or future) act ions and decisions, to justify them,and to suffer punishment in the case of eventual misconduct. In experiencesof political accountability, usually all three dimensions-information, jus ti-fication, and punishment-are present. However, they do not form a core ofbinary defining characteristics that are either present or absent and thatmust be present in all instances we describe as exercises of accountability.They are continuous var iables that show up to different degrees, with vary-ing mixes and emphases. Furthermore, even if one or two of them are miss-ing we may stil llegit imately speak of acts of accountability.To begin with, sanctions form an aspect of accountability many consid-er to be indispensable. Yet some agencies of restraint come to equateaccountability fundamentally with answerability. For ins tance, a good num-ber of valuable agencies of accountability, such as the Chilean and theSouth African truth commissions set up to inves tigate human rights viola-tions under predecessor regimes, have relied on only a soft form of pun-ishment-namely, the public exposure of criminal action .Furthermore, contemporary discussions on the autonomy versusaccountability of central banks (e.g., with respect to the fledgling EuropeanCentral Bank) reveal ideas of accountabil ity that are entirely detached fromnotions of punishment. Central bankers, if they accept the idea of account-ability at all, eguate it to the mere need to publicly explain their decisionsaft r th fa t. In n ral, m st expressions of vertical accountability

ConceptualizingAccountability 17

quences will usua lly appear as weak, toothless, diminished forms ofaccountability. They will be regarded as acts of window dressing ratherthan real restraints on power.The tight coupling of accountability and the availability of sanctionsreflect neoinstitutionalist common sense about the conditions of effectiveinstitutions. According to much of the new institutionalist literature, forrules to be effective they must be accompanied by mechanisms of monitor-ing that prevent the eventua l violation of rules from going unnoticed ( theinformational function of accountability). But they must also count withmechanisms of enforcement that get the incentives right by keeping actsof cheating from going unpunished (the enforcement dimension of account-ability).'Academic writers are often quite emphatic in stat ing that the capacityto punish forms an integral part of polit ical accountability.t- But poli t ica lactors, too, usually have a very keen sense for the pivotal importance ofeffective enforcement mechanisms that will enable agencies of accountabil-i ty to act forcefully. For example, in Taiwan, the Control Yuan (a branch ofgovernment that is charged with investigating the conduct of public offi-cials ) is able to impeach a public official-in effect, to indict and referthe official to the lega l authorit ies for further investigation and possibleprosecution. But because the ministry of justice is controlled by the rulingparty, this means in practi ce that government officials (especial ly high-ranking ones) are rarely seriously punished. The one structural change thatsenior off icials of the Control Yuan most want is the independent authorityto prosecute and impose sanctions on offending officials.PSimilarly, during Nigeria's Second Republic (1979-1983) an elaboratecode of conduct required public off icials to report their assets. ACode ofConduct Bureau was se t up to investigate the authentici ty of these reportsas well as charges of bribery and other wrongdoing. However, the Code ofConduct Bureau lacked the staffing to adequately inves tigate violations ofthe code by public officials, and the parallel Code of Conduct Tribunalnever sat (because the nationallegis lature never enacted the enabling legis -lat ion). Because polit icians feared no consequences for corrupt behavior,gross abuses of public offices, such as br ibery, embezzlement, extortion,and vote buying, quickly escalated out of control and generated the publicdis il lusionment with democracy that helped bring about the mili tary coup. ofDecember 1983.14Concerning the nature of possible sanctions, in the world of politics,the destruction of reputation through public exposure represents one of themain tools of accountability. But the pool of sanctions is rnuch wider. Theseverity of the sanc tion depends n th S v ri ty f th ff ns . In p liti cs ,r m val from ffi oft n J pr S nts th nl0HI QIlHIic ons qu n f mis-

n d u t. 01 1 1 in s t r u r nt H , Illh ll - 1 1 unt d i m i s 111. Ip P /11 IN n t i r Iy

appropriate punishments for a wide range of bad behavior. But as soon asactors violate legal dispositions, neither public criticism nor job lossappears sufficient. I llegal behavior, such as corruption or human rights vio-lations, calls for appropriate legal sanctions. When presumptive agents oflegal accountability, such as the Mexican National Human RightsCommission, for example, do not have prosecutory powers but can onlyissue nonbinding statements of an advisory nature, many observers will feelthat the soft and quasi-volunta ry version of accountability light theypracti ce does not deserve it s name and that, in essence, inconsequentialaccountabil ity is no accountabil ity at all . I f a police officer kil ls someone incustody without due cause and stil l walks free, i t does not sat is fy the princi-ple of accountabil ity if a journalist documents this abuse of authori ty or if ahuman rights ombudsman recommends that the officer be arrested andstand trial. Un les s there i s some punishment for demonstrated abuses ofauthority, there is no rule of law and no accountability.

Radial Concept


7/12

Conceptual and Normative Issuesnotion of poli t ic al accountability enters a world of power. Agencies ofaccountability strive to keep power from running wild; they strive tobound, to discipline, to restrain it. Their mission is to make power pre-dictable by lirniting its arbitrariness and to prevent or redress the abuse ofpower in order to keep it s exerci se in line wi th certain preestabli shed rulesand procedures. The existence of power provides thei r very raison d'tre.Without power, without the capaci ty to make decisions and the correspond-ing capacity to attribute decisions, it does not make any sense to talk aboutaccountability. Nobody can hold anybody accountable for things beyondthat person's control. I cannot hold you accountable for the bad weather(unless I subscribe to prescientific meteorological theories), nor can I holdyou accountable for the wrong weather forecas t (unless you are working inthe meteorological office).We rnight al so say that accountability presupposes personal responsi-bility. The two terms look very similar and are often treated as close syn-onyms. Yet if one equates accountability with responsibility someone,one has stil l to concede that there is no perfect overlap with responsibilityfor something. In fact, the lat ter presupposes the former, but the inverse isnot true. Accordingly, responsibili ty for something may go entirely withoutresponsibility to somebody. The premodern absolute ruler could be (andcould admit to be) responsible for lots of things but still reject having torespond to anybody (but God and history). In modern times, the presump-tive burden of responsibility has become an easy formula to legitimatepoli tical power (equiva lent to the idea of risk taking in econornic l ife). Andit has also become a formula to shed off prospective accountability: I amthe one responsible so I make decisions based on the way I feel I should; Ineed not tell anybody about those decisions before I make them; nor do Ihave to respond to anybody except to my personal conscience. In essence,while accountability force s power to enter into a dialogue, the notion ofresponsibility permits it to remain silent. While accountability builds on themodern idea that power and knowledge are separate goods, the notion ofresponsibility al lows powerful ac tors to maintain the illusion that theyknow what they are doing and therefore to di smiss irri tating questions thatdo nothing but dis turb their solemn and responsible exercise of power.I f accountabil ity indeed addresses agents who hold power and disposeof some margin of discretion, it follows that accountabil ity should not beconfused with narrow notions of regulation, control, or steering. Holdingpower accountable does not imply determining the way it is exercised; nei-ther does it aim at eliminating discretion through stringent bureaucratic reg-ulation. It is a more modest project that admits that politics is ahumanenterprise who e elements of agency, freedom, indeterminacy, and uncer-tainty ar ineradi abl ; that p wer cannot be ubject to full control in theSlri t,1 hni al s ns of rh w rd; and that cven in hypothetical world ofI rf '1 ti' '( untuhllhy, p< liti id pow r would .oruinu 1 pr duc harrn,

Conceptual iz in g Accountability 19

wh r m a ss media and civil associations act as accounting agencies rely onp u r Iy dis ursive forms of contestat ion and constraint , with no possibili ty01'111plyin negative sanctions other than publc disapproval.'>

sume way accountabil ity may be divorced from sanctions, in some'us S it rna y be an exclusive matter of sanctions. Accountabili ty understoodliS n r r ement may be cut off from questions of answerability. Thelnd n sian students who demanded the demise of General Suharto in early1 98 did not request further informat ion, nor did they desire to hear anydiscourses defending the long-reigning president. Or, less dramatically, theidea of electoral accountability--of voters holding politicians accountableat periodic elections-is also primarily related to the possibility of punish-ing pas t behavior by throwing the rascals out (even if between electionsincumbents may continually disclose their actions and justify them).Whereas in the first case, an outrageous as well as transparent violation ofsocial norms makes any addit ional informat ion superfluous and any addi-tional justification useless, in the second case, the communicative asyrnme-tri es be tween elit e actors and mas s publics render any dialogic idea ofaccountability impractical.The above examples illustra te tha t i t i s possible, in principle, to findinstances where the idea of accountabili ty is dissociated from one of itscore dimensions-be it enforcement or answerabili ty-without necessarilycreating diminished subtypes of accountabil ity as a result.ls 1concludeaccordingly that accountability does not represent a classical concept dis-playing a hard core of invariable basic characteristics. Instead, it must beregarded as a radial concept whose subtypes or secondary expres-sions do not share a common core but lack one or more e lements that char-acterize the prototypical primary category.?

A Modest Concept

Polit ical accountabil ity may be a broad and comprehensive concept. But itis also a modest concept. In part, its modesty stems from i ts potent ial one-dimensionality. As argued above, certain instances of accountabili ty do notinclude aspects of answerability, while others go without elements ofenforcement. But more fundamentally , polt ical accountabil ity is a modest. enterprise insofar as it accepts the realty of power as well as the relativeopacity of power.Th e Uncertai nties o f Pow erT hc uut lin i 1 n of p lili ni u Olll1tl1bilily iN to 011 11 ' 1 poliri ni pow r,noi to llmlnnl l t . In lh. 11 ,polllld i\ (lI1IlWIIIlI pl ' IH IJ ~)()N Hpow 1,I /{ II t'1I\ l1l h Id 0 I 1 1 l lilop \ 0 1 1OW 11 t i 111 II1 11 , W 1 1 1 1 11 IIW II 1 11


8/12

o Conceptual and Normative Issues

unrability, f ours ,falls utside th, nc rn d with po llt] a l ti untability

ConceptualizingAccountability 21

waste, and any other kind of irreversible pub lic bads that even idealagents of accountability could only ascertain, expose, and punish but nei-ther repair nor undo.Agents of accountability want to reach (partial) control over politicaldecisionmakers. If they already were in (fu11) control, their mission wouldmake no sense any more. I f 1control somebody, there is no point of makingthat person accountable to me. Accountable for what? For the things 1induce him or her to do? Robots that assemble automobile parts in aChrysler factory cannot be considered plausible addressees of accountabili-ty. The same applies to government officials who carry out decrees. Theyare supposed to do their work we11(and may be held accountable for it), butthey are not accountable for the rules they apply. Accountability concernsagents, not subjects . I t concerns those who exercise power, not those whoare subordinate to it. Or, more precisely, i t concerns subjects only as far aswe ascribe some degree of freedom to them. In an analogous way, i t con-cerns public employees only as far as we envision administrative organiza-tions not as mechanical conveyor belts of decis ions from top to bottom, butrather as loci of decisionmaking at a11hierarchicallevels.

self-contained arrogance of power. At the same time, however, legitimaterealms of secrecy exist where decisionmakers deliberate and decide behindclosed doors. We may think, for example, of cent ral bankers settingexchange rates or court judges forrning their judgments. Subjecting suchzones of legitimate confidential ity to procedures of accountability shouldnot open them to the eyes of the general public from beginning to end. Butit should oblige the involved participants to give ex post explanations andreveal the grounds for their decisions to the publicoWe should add that usua11y if i t wants to be effective and perceived assuch, democratic accountability must be publico As a rule, only publicaccountability can achieve its aim of curbing power, while confidentialaccountability, exercised behind closed doors, tends to be perceived as afarce, a car icature of accountability. For example, Soglo, former presidentof Benin, disclosing his personal assets secretly-? hardly qua li fied him as achampion of accountability.This publicity principle applies not only to accountable ac tors butalso to accounting agencies. This makes accountabil ity dis tinct f rom super-vision, for instance. The supervisor may remain in the dark, the unseen eye.The agent of accountability must come to the forefront and open itself tosecond-order observation: the observer observed. In this sense, institutionsof accountability are vampires in reverse. They can live only as long asthey act in the daylight of the public sphere, and they crumble and die assoon as they enter the shadows of privacy and secrecy.s?The public nature of accountability serves all three aspects of account-ability: information, reasoning, and punishment. It does more than bringthe forceless force of the better argument21 upon the conduct of theaccountable party. It al so involves an important form of sanc tion. It exposescases of misconduct to public opinion, which often provokes highly damag-ing reputational consequences.P

T h e O p a ci ty o f PowerAccountability as answerability aims at creating transparency. By demand-ing information as we11as justification, it wants to shed light into the blackbox of politics. In this it is similar to notions such as oversight, supervision,and monitoring. But again its project is less ambitious. Agents of account-ability do not pretend to supervise everything. In fact, they assume thatnothing close to close oversight is taking place and accept that their gen-uine field of competence consists of unobserved and often unobservableactions.P In this sense, accountability (as answerability) presupposesimperfect information. If the exercise of power were transparent, therewould be no need for holding anybody accountable. The demand foraccountability (as answerability) originates from the opacity of power. In aworld of perfec t information, it would be pointless to ask political actorswhat they have been doing, or intend to do, and why. We would alreadyknow. In Jeremy Bentham's Panopticon, any pretense to exercise account-ability is void. From i ts high center, we can see everything and interveneanytime. We do not have to ask anything. We just watch and punish. In thereal political world, however, most things are not accessible to direct obser-vation. And even if informational barriers to the political process did notexist, we could s ti11not look into the future (which creates the need for ret-rospective accountability) , nor could we look into decisionmakers' heads(which creates the need to ask for justifications).When agents of accountability intrude into the op aqu e rea lrn s fpower, they are concerned with the classical arcanae imperii, thc dark play-grounds of power shut ff from publi scrutiny f r no th r r nson (han Ih

Who Accountable for What?In a11walks of life we can hold people accountable for a11kinds of things.In our private lives we wade in a constant stream of accountabili ty initia-t ives. People hold their children, parents , par tners, friends, neighbors, col-leagues, and fellow citizens accountable for any kind of presumptive mis-behavior-for political incorrectness, insubordination, disorder liness , badmernory, drinking and smoking, sexual misconduct, sinful behavior, lack ofcourtesy, s trategic errors, factual ignorance, whatever. Because there arelots of rules that guide our private lives, there are lots of opportunities forprivate agents of accountability to step in to monitor and enforce compli-anc .Th vs r 01111 r priva r alu i d i '11 01 1 I 1 1 11 .I (lO whl 'h Is


9/12

22 Conceptua l and Normative Issuesability rely on different enforcement mechanisms. The responsibility forexercising both political and moral accountability applies to citizens, civilassociations, mass media, and opposition parties. Both administrative andfinancial accountability are often entrusted to specialized agencies, such asombudsmen, administrative courts, accounting offices, and anticorruptionorganizations. Ethics commissions and disciplinary courts are commoninstitutions of professional accountability. And judicial systems are incharge of guaranteeing legal as well as constitutional accountability. As thisquick list ing makes clear, distinguishing between different accounting par-ties constitutes a useful basis for constructing further subtypes of account-ability. But we may easily drown in a sea of innumerable agents of account-ability unless we drop some conceptual anchor that puts order into thepotential proli feration of subtypes.w Guill ermo O'Donnell's seminal dis-tinction between vertical and horizontal accountability provides, I think,such a heuristic ordering device.s?The notion of vertical versus horizontal relations of accountabilityparts from the conventional spatial metaphor of power, the classic image ofpyramidal hiera rchies: he ight correlates with resources, above equalspower, below equals powerlessness. In this sense, vertical accountabilitydescribes a relationship between unequals: it refers to some powerfulsuperior actor holding some less powerful inferior actor accountable.Or vice versa The concept is indeterminate in this crucia l aspect of direc-tionality. It leaves open whether accountabil ity is meant to flow top downor bottom up, whether it forms part of the normal exercise of power orwhether it implies an inversion of usual power relations. The classicinstance of the former is bureaucratic accountability, in which higher-rank-ing public officials ( principals ) try to control their lower-ranking subor-dinates ( agents). In representat ive democracies, the most important vari-ant of the latter is electoral accountability, in which citizens hold judgmentover their representatives through periodic elections.28By contrast, horizontal accountability, taken literally, describes a rela-tionship between equals: it refers to somebody holding someone else ofroughly equal power accountable. In democratic theory, the division ofpower-the executive, legislative, and judiciary constraining each otherthrough the classic checks and balances-represents its prototypicalexpression. Yet demanding a rough equality of power for horizontal rela-tions of accountability establishes a tough crit erion, an overly tough one .Power is a (relational) property that is hard to measure, and trying to identi-fy instances of roughly equal political power in the real world of democrat-ic politics is probably an impossible mission.In addition, an accounting party, if it is to be serious, cannot stand onequal footing with the accountable party. At least in its sphere of compe-tcncc, it must be cvcn more powerful. And certainly within the boundariesof ils [urladtcrlcn, which m n y be vcry narrow, it must be, as O'Donncll p u e s

Conceptualizing Accountability 23in its wide sense-with acts of accountability addressed to public officials(the whole personne l employed by the modern state: politicians, civil ser-vants, judges, police officers, military officers, diplomats, etc.). In its nar-row sense, the term political accountability covers the activities of only theprofessional category listed f irst : politicians, such as cabinet members, leg-islators, and party leaders.Parting from the broad concept of political accountability, which con-cerns the behavior of any public official, one might introduce a number offurther subtypes according to the polit ical subsystems that are subject toaccountability. The referents and the labels of these subtypes will dependon the analytic distinctions one chooses to draw. A conventional way ofslic ing the modern pol it ical world would produce at least the followingbroad categories: governmental , legislat ive, bureaucratic, judicial, and mili-tary accountability.Looking at the addressees of accountability (who is accountable?) isone way of ordering the wide f ield of poli t ical accountability. Inquiring intothe criteria of accountability (accountable for what?) is another way. So, ifwe hold public offi cials accountable, what do we hold them accountablefor? What yardsticks do we use? How do we define good and bad conductin office? How do we accordingly define the abuse of off ice that the exer-cise of polit ical accountabil ity s tr ives to prevent and punish?It is obvious that we may hold public off icials accountable for lots ofthings and that our acts of accountability may be inspired by many poten-tially conflicting principles. A quick search through different standard s ofaccountability suggests tha t we should distinguish at least the followingvarieties of political accountabili ty (in its broad sense) : polit ical account-abili ty (in i ts narrow sense) assesses the appropriateness ofboth substantivepolicies and policymaking processes, but it also brings judgment on thepersonal qualities of political actors; administrative accountability reviewsthe expediency and procedural correctness of bureaucratic acts; profession-al accountabili ty watches over ethical standards of professionalism, such asmedical, academic, and judicial professionalism;23 financial accountabilitysubjects the use of public money by state officials to norms of austerity,efficiency, and propriety; moral accountability evaluates political acts onthe basis of prevailing normative standard s (independent of formal rulesand regulations); legal accountabi lity monitors the observance of legal. rules;24 and constitutional accountability evaluates whether legislative actsare in accordance with constitutional rules.25

Accountable to Whom?Who are the agents of political a cco un ta biti ty ? Who tire us sociu carriersa nd w hut a re its in s jllu jionul fo 1 n1 fl 1 1dop cnds. I )lfTt lrtlll1 I 'O IlIlH )ji ucc ount-


10/12

24 Conceptual and Normative Issues

When agents of accountability develop into powerful actors, when their cri-teria f judgrn nt ar controversial, and when it is difficult to monitor theirp rf rrnan w Ia th pr bl m l' S nd- rder a e unta ility: h w ca nw hold inSlilllliolls of 11' iounrnhility fI ountubl th ms lv s? This prob-

ConceptualizingAccountability 2S

journalist denouncing nepotism within the bar associat ion represent s anexample of horizontal accountabil ity within civil socie ty; and the chamberof cornmerce questioning the constitutional ity of se lective price subsidiesmay count as an example of horizonta l accountabil ity from civil socie ty tostate.32The controversial question is, however, whether we shouldreservethe notion of horizontal accountability to intrastate relations, asO'Donnell proposes, or else also include in the category civil societyactors that hold state agencies accountable, as authors such as RichardSklar, Philippe Schmitter, David Stark, and Lszl Brusz t suggest.ttBoth conceptual izations of horizon tal accountabi li ty seem quite leg iti -mate. The main objection against the second option of extending thenotion of horizontal accountabil ity to civil society-state relat ions res idesin the structural asymmetry between state and nonstate actors. Civilassoc ia tions may represent st rong factual powers poderes fcticos).But they can never match the state 's monopoly of legitimate physica lviolence or its privileged position as the source of law-of uniform, cen-trally defined, collect ive ly binding rules. Thus, speaking of horizontalrelations between the two spheres always sounds like something of aeuphemism. In addition, ascribing one clear meaning to the notion ofhorizontal accountabil ity carri es the benefit of avoiding conceptualambivalence. Accordingly, most authors in this book follow O'Donnell,insofar as they assume that all cases where civil society actors (citizens,civil associations, or mas s media) try to hold state agents accountablefall into the category of vertical accountability, reserving the notion ofhorizontal accountability to all acts of accountability that take placebetween independent state agencies.As Robert Pastor (in this book) has forcefully reminded us, a ll thenotions of accountability discussed so far share one basic presupposition:they take for granted that polit ical accountabil ity works within the confine-ments of national political systems. As a consequence, they overlook inter-national actors (governmental as well as nongovernmental ones) as poss ibleagents of accountabil ity. Since these exte rnal actor s do not fi t easi ly intothe vertical/horizontal dichotomy constructed along the distinction betweenstate and civil society, they represent , as Pas tor puts i t, a third dimensionof accountability.

it (in this book), legally enabled and empowered to impose sanctions onthe accountable par ty, to punish improper behavior on its behalf. Relaxingthe defining crit erion of rough equal ity in order to replace it by some loosercategory, such as an approximate equivalence of power.t? does notchange this paradox of horizontal accountability, of specialized agenciesholding actors accountable who are immensely more powerful on allaccounts except in the agency's speci fic sphere of competence . Besides,accountabil ity as answerabi li ty does not even include the power to punishbut only the right to get an answer. In terms of power, it may take placeunder extremely asyrnmetric conditions.Some authors take these methodological and concep tual diffi cul tie sinto account by complementing the binary spat ial metaphor of vert ica lityversus horizontali ty with some intermediate category, such as diagonalor oblique accountabil ity.? This solution, however , does lit tle to clari-fy the mat ter. How should one determine, in prec ise terms, inte rmedia teangles of inclination between horizontal and vertical lines of account-ability?If indeed in relations of horizontal accountability the agent and theaddressee of accountabil ity need not possess comparable power resourcesbut in fact may represent actors of gross ly unequal powers, it does not seemhelpful to comprehend horizontality in a literal way as a relationshipbetween actors of equal or equivalent power. Rather, i t seems more produc-t ive and of more genera l applicabi lity to st ipulate tha t the accounting partymust be independent from the accountable party in a11dec isions that con-cern its field of competence. Ideally, both parties form relativelyautonomous powers31 that do not stand in a re lat ion of formal subordina-t ion or superiori ty to each other. In other words, horizontal accountabilitypresupposes a prior divis ion of powers, a certain interna functional differ-entiation of the state.Defining horizonta lity in terms of autonomy resolves one main con-cep tual problem that has accompanied the concept of horizontal accoun t-abi li ty. Another problem concerns the concept 's basic domain of applica-tion: where are the autonomous agents of horizontal accountabi lity to belocated? Just in the state? Or also in civil society? It is clear that verticalaccountabil ity may take place within the s tate as well as within civil societyand across the boundaries of the two spheres. Tax off icials who stage sit- insin front of the finance ministe r's office are an example of vert ica l account -abil ity within the state ; union members who accuse the ir long-standingleaders of corrupt behavior are an example of ver tical accountabil ity withinc ivi l society; and cit izens who vote incumbent parties out of office are anexample of ver tical accountability running from civil society to state. Inprincipl e, horizontal accountability to m ay tak pla b th within andacross state and civil ciety. A jud m lIin a rni li tary ffi J t sti-r y in urt is an xarnpl r h riz nLUI o u n t u b l l it y wilhin 111 st 11; f1

Second Order Accountability


11/12

Conclusion

Conceptualizing Accountability 276 Conceptual and Normative Issueslem seems parti cularly pressing with the kind of inst itutions discussed inthis book: specialized, nonelective, autonomous state organizations that aresupposed to pursue their narrow missions with professional single-minded-ness, while they are nevertheless vulnerable to ineffic iency and abuse thesame as any other locus of power.v 'Of course, one may reformulate the challenge of second-order account-abil ity as a chal lenge of n-order accountabil ity. Since any second layer ofins ti tutional accountabil ity is susceptible to the same kind of fai lures as thefirst layer, we face the possibil ity of an in fini te r egress. There are two waysto avoid it. We may establish reciprocal accountabil ity: two agents, A andB check and balance each o ther . Or we may establi sh recursive account-abi li ty: A is accountable to B, who is accountable to e, who is accountableto A again. It is hard to imagine that, at least in the long run, insulated andisolated ins ti tutions of accountabil ity can survive unless they are embeddedin such recursive cyc1es or sys tems of accountabil ity. As O'Donnell puts it( in this book) , achieving a s ignificant degree of . .. accountabil ity requiresthe coordination of several agencies, each of them subject to divide etimpera strategies.

This review of the concept's semantic field and internal structureshould establi sh a re liable basi s for the other chapters in the book, even ifnot al l authors adhere to exact ly the same interpre tat ion of accountabil ityand i ts subtypes. In more general t erms, i t should also provide useful ana-lyt ical tools for understanding and advancing academic as well as polit icaldi scussions on poli tic al accountabi lity-this obscure object of poli ti caldesire and institutional designo

Notes

Where did our surveying expedition into the structure of meaning of theconcep t o f pol iti ca l accountabil ity take us? Let us sum up the main coordi-nates of our conceptual journey. Poli tical accountabil ity, we stipulated, rep-resents a broad, two-dimensional concept that denotes both answerability-the obligation of public officials to inform about their activities and tojust ify them-and enforcement -the capaci ty to impose negat ive ~anc tionson officeholders who violate certain rules of conduct. In expenences ofpolit ical accountabil ity, both aspects are usually present . Y~t.instances existwhere e ither the call for sanc tions dominates (accountabi li ty as enforce-ment) or the quest for information and justification (accountability asanswerabi li ty). Given that the not ion of accountab ili ty i s no t buil t on theillusion that power is subject to full control and can be opened up to fulltransparency, but rather accepts and addresses the uncer tai~ty. and opaci~yof power, we character ized it as a modest concept. After clarifying the bas icmean ing of accoun tab ili ty, we stepped into the vast fie ld of di fferent sub-types of accountabil ity. We sketched some ways of dis tinguishing differentforms of accountab ili ty, accord ing to it s addressees, its criteria, and itsagents. Given the thematic focus of the book, our discussion centered onthe distinction between vertical and horizontal accountability. It defined thesomewhat vague and metaphorical notion of horizontali ty as a relation-ship between state agencies that must possess a certain degree of mutualindependence, rather than equivalent power resources.

In a cademic life, it r epresen ts a rar e and p reciou s event when an autho r r eceivesseveral pages of detai led, sympathet ic , and const ruct ive feedback on a dra ft . 1 thankan anonymous r eviewer and Larry Diamond fo r writing such comments. Withoutthe ir insightful and often fundamental c ritic ism, thi s chapter would be very dif fe rentf rom what it is now. My s incere thanks to both of them.1. The Federalist Papers, no. 51.2. For a brief historical review of moral and institutional solutions to theproblem of corrupt power , s ee Fontana , The Roots of a Long Tradi tion .3 . Richard Sk lar even asserts th at th e norm of accountab ility appear s to bethe most widely practiced of democratic pr inciples (Sklar, DevelopmentalDemocracy, p. 714).4. Plattner, Comments, p. 1.5. Hickok, Accountability, p. 9. See also Sklar, this volume.6. See Plattner, Cornments , and Sklar, this volume. This twofold answer-ability refers to past as well as to future acts. Especially in poli tic a l contexts,accountable agent s have to s tand up not only for wha t they have done ( re trospectiveor e x post accoun ta bility) but also f or what they plan to do (p rospec tiv e or ex an teaccountability). See also Maravall, Accountability,' and Schmitter, this volume.7 . S ta rk and Brusz t, Postsocialist Pathways, define accountabil ity as moni-toring (p. 195), thus equating it with the informational dimension. Yet , a t o therplaces, they use the term accountability interchangeably with deliberat ion (p. 189),thus equating it with the dialogic dimension.8. Bennett , The Paradox of Public Discourse, p. 794.9 . S ta rk and Bruszt, Postsocialist Pathways, emphasize the connectednes s ofboth aspec ts of account s. They develop the metaphor of democr atic politiciansopening account s, asking voters for c redi t, and giving account s in order to maintaintheir creditworthiness in an origina l and ins ight ful way (see pp. 192-196).10. See O'Donnell, Delegative Democracy, p. 61.11. See, e.g., Ostrom, Governing the Commons.12 . See, e.g., Collier, O'Donnell, Schmitter, and Sklar (this volume); Maravall,Accountability.13. See Diamond, How Well Is Taiwan's Democr acy Doing?14. See Diamond, Political Corruption.15. On the distinction between hor iz ontal and vertical accountability, seebelow.16. On dirninished subtypes, see Collier and Levitsky, Democracy withAdj ctives.

17. n I tlR Si al und radial onc pts, s ollier and Mahn, Conceptual'A l' t 'hin 'R vlsl; u.


12/12

28 Conceptual and Normative Issues

18. The preceding reflections do not apply for accountability as enforce-ment. As we noted above, rule violations may already be in the open as publicfacts that are not opaque; nei ther does the punishment of those violations add to thetransparency of politics.19. As reported by Galtung, Developing Agencies of Restraint, p. 11.20. Of course, there are exceptions to the general rule of publicity. Thebureaucratic accountability of subordinates to superiors usually remains an inter-nal affair wi th in the administrative organization concemed. And in some traditionalrealms of state secrecy, such as foreign intelligence services, accountability, if ittakes place a t all, does so in a confiden tial way, closed off to the eyes of the generalpublico21. See, e.g., Habermas, Theorie, p. 47.22. See, e.g., Pizzomo, Representation, p. 8.23. Much of professional accountability pertains to the sphere of civil soci-ety . At the same time, the category represents a weighty aspect of the accountabilityof public officials, and the more so, of course, the larger the public sector in a givencountry.24. Note that the last three categories-moral, financial, and legal account-ability-represent entirely crosscutting forms of polit ical accountabili ty tha t maystrike any actor in the political system, be it a party leader, the head of govemment,a member of parliament, a judge, a simple state official, or whomever.25. This volume basically covers the last two forr ns of accountability: legaland constitutional.

26. The picture would look even more complicated if we were to includeforms of accountability not only to concrete actors but also to ideal, subjectless cat-egories: to history (historical accoun tability ), to our personal consc ience (super-egoaccountability), to God in heaven (ce lest ial accountability) , or to pastgenerations(ancestral accountability). 1 owe the term celestial or heavenly accountability toRichard Sklar (this volume).27. See O'Donnell, Delegative Democracy. For a similar early conceptual-ization under the heading of lateral accountability, see Sklar, DevelopmentalDemocracy.28. Both types of relationship may be modeled along the lines of principal-agent theory (see, e.g., Maravall, Accountability; Moe, Poli t ica l Institutions,pp. 232-233; Roeder, Real Sunset, chapte r 2 , pp. 22-40).29. See, e.g., Schmitter, this volume.30. See also Schmitter, this volume, p. 62, note 4.31. O'Donnell, Delegative Democracy. See also Diamond, The End, pp.8-9. 32. For an insigh tful discussion of the many forms, sites, and protagonists ofaccountability, within the state as wel l as within civil society, see Sk lar, Develop-mental Democracy.33. See the contributions of O'Donnell, Sklar, and Schmitter, this volume, aswell as O'Donnell, Delegative Democr acy, Sklar, Developmental Demo-cracy, and Stark and Bruszt, Postsocialist Pathways.34. For a refreshing recognition of the need for second-order accountabilitywith relation to central banks, against the prevailing mainstrearn that favors maxi-mizing unqualified central bank independence, see Johnson, this volurne, Schrz,Independence Versus Accountability, and Whitehead, Models f Central.Banking.

3Horizontal Accountability

in New DemocraciesGUILLERMOQ DONNELL

For Gabriela,my beloved agency o/ horizontal accountability

My interest in what 1 .labe.led h.orizontal accountabilityl stems from itsabsence. M.a~y countnes, . ll l Latn America and elsewhere, have recentlybecome pol~t~cal de~ocracles, or polyarchies. By thi s 1mean tha t they satis-fy ~he ~OndltlOns slp~~ated by Robert Dahl for defining this regime type.?SatlSf~lll~ these conditons is no mean feat: some countries continue underauthont~~ rule, a~d others, even though they hold elections, do not satisfythe condlt~ons o~fair and free competi tion stipulated by the defini tion ofpolyarchy. In this cha~ter ~do not deal wi th the latt er cases; my focus is onth?se that ar~ polyarchies, the sense just defined, but have weak or inte r-mltt e~t honzontal accountability. This refers to almost all the LatinAmenc.an countries, including some rather old polyarchies such asColombl~ and Venezuela+ This category also pertains to some new Asian~olyar~hies, such as the Philippines, South Korea, and Taiwan, as well as,mcreas.mgly so, .an old one, India, and some of the postcornmunist countriesthat mlg~t q~ahfy as polyarchies-Russia, Belorussia, Croatia, Slovakia,and Ukram~, .a.ndperhaps also Hungary, Poland, and the Czech Republc.sBy d~f~mtlO~, In these countries the electoral dimension of vertical~ccount~~lhty exists. Through the means of reasonably fair and free elec-tions, ciuzens ca~ punish or reward incumbents by voting for or againstt~em, or the candidates ~h~y endorse, in the next election. Also by defini-non, the fre.edoms of opiruon and of associaton, as well as access to rea-sonably varied sources of information, permit articulating demands to andventllall~ dcnoun in wr n fui acts f, publi auth rities, This is h~lped

by 111 XINt n of r usolllbly fr m da, Bis I mand d l Y Ih d finili n

schedler 1999 conceptualizing accountability

Documents