Vous êtes sur la page 1sur 8

Mandates or Mandarins?

Control and Discretion in the Modern Administrative State

Author(s): Joel D. Aberbach and Bert A. Rockman
Reviewed work(s):
Source: Public Administration Review, Vol. 48, No. 2 (Mar. - Apr., 1988), pp. 606-612
Published by: Blackwell Publishing on behalf of the American Society for Public Administration
Stable URL: http://www.jstor.org/stable/975761 .
Accessed: 09/02/2012 18:51
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Blackwell Publishing and American Society for Public Administration are collaborating with JSTOR to
digitize, preserve and extend access to Public Administration Review.


Joel D. Aberbach, University of California, Los Angeles
Bert A. Rockman, University of Pittsburgh
The developmentof the administrativestate and the
growth of political *democracyconstitute two of the
most distinctivetendenciesof moderngovernment.The
developmentof an advancedadministrativeapparatus
carrieswith it claimsto the valuesof continuity,professionalism, expertise, and effectiveness. The other
development,that of politicaldemocracy,encompasses
claims to the values of responsiveness,direction, and
revitalization.Notwithstandingthe desirabilityof each
set of values, the meansfor meshingthem in an optimal
mix are hardly obvious. Even though it is widely accepted in democraticsettings that the permanentadministrationmust be accountable to constitutionally
electedor delegatedpoliticaloverseers,the preciseterms
of this agreementare much more controversial.
Almost certainly,few of us come to see the struggle
betweenpolitical control and administrativediscretion
in entirelyneutralterms. Typically, dependingon our
particularinclination, we tend to adopt perspectives
that place more weight either on "political" or on
"administrative"values, regardlessof the importance
we attach to the need for an optimalmix.
Partisans of political leadership(and these almost
alwaysincludethe incumbentset of leaders)are doers,
not doubters.Theywanttools, not obstacles.To the extent that doubt existsabout the willingnessof careeradministratorsto carryout faithfullythe policy directions
of the political leadership, career administratorsare
viewed by political actors as impedimentsratherthan
implements. Partisans of politics, consequently,typically look to enhanceproceduresfor controland supervision of the permanentadministrativeapparatusand,
when deemednecessary,to politicizeit.
Partisansof the careeradministration,on the other
hand, view it as the ballast that maintainsthe ship of
state in unsteadyseas. Its resistor-likequalitiesto the
super-chargedenthusiasmsof new political leadersare
seen as a virtue, not a vice-a deterrent,in fact, to
longer-run damage inflicted by political leaders on
themselvesas well as on the organizationalfabric of
government. Partisans of public administrationthus
decry efforts to reduce the independenceof career
officialdom or to restrict severely administrative
The political leadershipview in the modern democratic polity is one that we characterizeas the "mandate" perspective.Underlyingit is the logic that the
electedpolitical authoritieshave eithera right, an obli-

* Although wide agreementexists that in democratic

settingsthepermanentadministrationmust be accountable to constitutionallyelected or delegatedpolitical
overseers,theprecisetermsof the agreementarecontroversial. The termsare especiallyelusive and unclearin
the United States because the separation of powers
system clouds any straightforwardprincipal-agentrelationship betweenpolitical authoritiesand career officials. Competitionbetween the political principals, indeed, makes the administrativeapparatusa resource
worthcompetingfor in an effort to influenceprogrammatic control overpublicpolicy. Some recentliterature
has emphasized the need for presidents to assert
stringentcontrol over the administrativeapparatusin
order to assurecompliancewith theirgoals. Politicization of the federal bureaucracyis justified therebyas
beingeitherin the broaderpublic interestor at least in a
president'sown interest. Weargue that neitheris likely
to be the case, and thatsuch efforts inviteretaliationon
the part of Congress.

gation, or a legitimateneed to pursue their goals and

policy proposalsand that it is essentialfor the operative
instrumentsof governmentto be in strict compliance
with these. The next step in this logic goes farther-indeed, a criticaldistance.The next step is that discretionaryauthoritywithinthe administrativeapparatuscan be
meted out only to those who meet requisitetests of ardor for the goals and methodsof the electedauthorities.
The administrativeview we shall characterizeas the
"mandarin" perspective-a term that resonates, for
historicalreasons, better in Europethan in the United
States. The essence of this view is surely applicableto
the Americansettingas well. It is that a professionalized
bureaucracy(which came late to the United States, we
should note) elevates the effectivenessof government.
The "good government" inclinations of the Progressives,for example,predisposedthem to what might
be calleda democraticmandarinate-the synergisticfusion of executive leadership from a democraticallyinspired elected executive and an efficiency-inspired
professional civil service. Historically in the United
States, much of the modern administrativeapparatus
was createdlargelyto advancethe goals of proficiency




and universalisticstandardssoughtby the Progressives,

and later it was used to advancethe goals of social and
economic reform and the developmentof the welfare
state throughthe New Deal, later fortifiedby the Great
Society. A high degree of congruencein purpose between the presidencyand the careerexecutivewas once
thoughtto exist-a truly democratic(but probablyalso
Democratic)mandarinatewas seen to be in the service
of the national interest(as that largelywas defined by
the president).
Althoughno presidentis ever preparedto leave what
he regardsas truly central activities to the career executive,the broadpremisesof whatpresidentsand their
administrativeapparatuswere about appearedto be in
general concordance. Well-articulatedand clear-cut
strategiesfor controllingthe administrativeapparatus
or cuttingit out of the action would awaitthe machinations of the Nixon WhiteHouse and its successors,most
notably, the Reagan Administration.What the Nixon
White House made clear in its operativepremiseabout
the bureaucracywas that it assumed noncompliance
ratherthan concordance.Moreover,it conceivedof the
Washingtonbureaucracyas tendingtowarduncontrollable fission ratherthan synergisticfusion. Whateverthe
realities of the situation, the underlyingattitudesand
perceptionsof the relevantactors have determinedthe
atmospherein which these relationshipsrecentlyhave
developed. The self-perceived possessors of the
democraticmandate worked to tighten the leash, to
diminishthe possibilitiesof noncompliantbureaucratic
tactics, and whereverpossible,to ensurethat implementation be carriedout only by trusted agents. The imperativeto commandhas grownincreasinglycompelling
from the perspectiveof the White House.

The IntellectualJustification
of PoliticalCommand
In the American case, however, the constitutional
basis of hierarchicalcommandis absent or, more properly, it is pluraland thus potentiallycontradictory.In
Richard Nathan's words, "it is the wonderfully
animated, competitive, and open character of the
Americanpolitical system that distinguishesit among
the democraciesof the Westernworld."'
It is exactly this competitiveness-a political market
systemas we shall think of it-that makesthe administrativeapparatusa resourceworthcompetingfor in an
effort to influence programmaticcontrol over federal
policy. A system of segmentedpower such as that exhibited in the syndromeof subgovernmentaldomination over programs(the triad of congressionalcommittee or subcommittee,clientelegroup, and bureau)produces what economists and, in their own way,
presidentssee as inefficientequilibria.2Whileeconomic
theoristsmight define these inefficientequilibriain the
form of misallocated resources, presidents tend to
define them in the form of subsystemicresistancesto
policy change.
In recentyears, the president'ssideof this problemMARCH/APRIL 1988

his abilityto managethe executivebranchand his need

to procure resources in the competitive struggle to
govern-has been voiced in sophisticatedways. Richard
Nathanarticulateswell the view that presidentsnot only
need to, but properlyought to, "influenceadministrative processesin a way that enables(them)to move forward on important policy objectives."' Clearly, it is
within the power of a presidentialadministrationand
within, broadly speaking, the norms of American
politics and governmentto make ideologicalharmony
an importantcriterionfor noncareeradministrativeappointments.The key obviously is how the "reds" interact with the "experts," and whether the "change
agents" recognize any legitimate bounds to their
strategiesfor effecting change. Above all, the central
issue is how the presidentialadministrationin its efforts
to influence administrativeprocesses interacts with
otherlegitimateauthorities,especiallyCongressand the

Even though it is widely accepted in democratic settings that the permanent administration must be accountable to constitutionally elected or delegated political overseers, the precise terms of this agreement
are much more controversial.
Nathan concludes, however, that because the
Americanpolitical system is dynamicand competitive,
"leadership is hard to exercise. . . . Policy changes are
not easy to achieve, yet are often needed. . . . [Conse-

quently, because] American national government at

high levels is not a subtle business . . . the administra-

tive strategyof the presidencyis a valid and valuableinstrumentof presidentialleadership.''1 In otherwords,it

is legitimate for presidents to seek to politicize the
bureaucracyon behalf of their goals becausepresidential leadershipis essentialto the system.Whenthe wheel
turns, other presidentswith different goals may also
legitimatelyseek to politicize the bureaucracyto their
own ends. The model is, as a formerpresidentused to
say, perfectly clear. It is collectively rational for the
system that presidentsshould command, and it is individuallyrationalfor presidentsto seek to command.
In an especially sophisticatedanalysis, Terry Moe
argues correspondinglythat a system such as that
describedby Nathan gives a rationalpresidentfew options.' Whetherindividualrationalityleads to collectivelyrationalsolutionsis a matterthat Moe leavesopen
to debate. Even though Moe seems stronglyto imply
that presidentialpoliticizationof the bureaucracy,includingthe institutionalpresidency,is a good, his argumentis couchedverymuchin the languageof individual
rationality.Whatis a rationalpresidentto do giventhe
logic prevailingbetweenincentivesand institutions?The
answerseemsto be to strivefor controlover everything
that is not nailed down.
Whetherpresidentialcommandis a good or a bad is
not Moe's fundamentalpoint. Presidentsseek to assert
control over what they can, he asserts,mostly because



they must. The maximization of control is viewed as a

systemically necessary strategy.
In the final analysis, writers as different as Lowi,
Rose, Nathan, and Moe all have bought into the mandate theory. Putting other analytic problems with such a
theory to one side, however, only the system of government that Rose discusses (British party government and
parliamentary supremacy) has institutions that are consistent with the premises of the mandate theory.6 In the
more structurally complex American system, Lowi has
chosen the statutory instrument as the anchor.7 This implies a kind of congressional supremacy even while it
promotes both administrative and political inflexibility.
Nathan and Moe, on the other hand, appear to gravitate
to the opposite pole, namely that executive command is
an appropriate (either desirable or simply necessary)
form of politicization. Yet, the theory of organizational
command and the theory that constitutionally organizes
the American system of government are at odds.8 The
point is that in the United States it is not enough to talk
about what politicians have a right to; one must specify
which politicians. That being said, an even more fundamental point about the American system that follows
from it is that in a system of divided authority, to say
that politicians have the right to control is not the
equivalent of saying that the president has the right to
control. Such rights, as Neustadt once noted, are joint
property rights.9 And, as Neustadt, in essence, also saw,
for such rights to be exercised, they would have to be
jointly authorized.-" It is possible, perhaps even probable, to suggest that this may be asking too much of a
system of divided authority and of a system that frequently also divides this authority along partisan lines.
But it also is likely that such a system requires either
unusual consensus-building skills and/or exceedingly
clear political signals from the electorate to alter existing
equilibria. Otherwise, presidents belatedly may come to
discover many adverse political effects from their efforts to monopolize a shared resource.

The PresidentialRole
in the AdministrativeProcess
More and more, however, what the White House
wants of civil servants, as ex-White House aide (and not
just coincidentally also ex-convict) John Ehrlichman so
picturesquely put it, is the following: "When we say
jump, the answer should be 'how high?"'
In recent decades, though, presidents and their entourages have come to conclude that when asked "to
jump," bureaucrats are not immediately inclined to ask
"how high?" but rather "to where?" For administrations bent on redefining the role of the state or just
simply jamming through their definition of priorities,
questions and conditionals are mere impediments. Accordingly, they conclude that it is best to cut the
operating agencies out of the action as much as possible
(centralization) and, when that is not possible, to cut the
careerists out of the sphere of potential influence while

relying on increased layers of politically faithful appointees(politicization).

The logic, as presidentsare inclinedto see it, is that
popular sovereigntyempowersthem to command the
apparatusof government.Even if one wereto conclude
that the only concreteexpressionthat could be given to
the public interest lies in the momentarywill of the
authorizedpoliticalleadership,the fundamentalflaw in
this conceptionis that this will is not derivablefrom a
single source. Membersof Congressalso lay claim to a
piece of the mandate. When the political will of Congress and the presidentare coincident, ironically, the
need for exclusivityof control over the administrative
apparatusdiminishes.When they are in conflict, it is
likely that exclusivity of claims for control will be
countered.It is certainlylikely that when institutional
interests clash and presumptive behavior increases,
nothing in Washingtonwill stay uncontestedfor long.
That includescontrol of the administrativeprocess.
Increasingly, it seems, presidents and political
theorists find the idea of "neutral competence" impossibleto describe.No one plausiblycan lack interests;
thus, all adviceor discretionarypossibilitiesare skewed.
The sentiment on behalf of politicization necessarily
assumesthis. Consequently,it follows that if all "parties" have interests, the concept of "neutral competence"lacksoperationalmeaning.If that is so, then it
is clear that the career executivesthemselveshave to
meet political criteriaor, as a group, be buried sufficiently far from the centers of power to preventthem
from exercisingmeaningful discretion or from being
able to influence decision makers. The decline of the
neutral competence ideal correspondsto the rise in
Washington of the adversarialideal-the belief that
everyone has an interest that they are seeking to optimizeand that all expressionsof collectiveor publicinterestare only facades(evenif theseareinternalized)for
the operationof individualinterestor preference.Accordingly,withoutpresidentialcontrolof the executive,
it is believedby many advocatesof presidentialcontrol
that the expressionof those interestsand preferences
will be chaotic overalland unaccountable.
The case for presidentialpoliticizationof the executive boils down to these suppositions.The presidentis
the supreme legitimate governor in the American
system. And since no one possibly can be neutral,it is
necessaryto assurethat the apparatusworksunequivocally on behalf of presidentialgoals and needs.

CollectiveRationality:Controlor Synthesis?
Politics provides energy and revitalization while
bureaucracybrings continuity, knowledge, and stability.11One can exist without the other but only to the
detriment of effective government. The problem for
governmentand, in our view, the public interestis not
to have one of these values completely dominate the
other, but to providea creativedialogueor synthesisbetween the two. In recenttimes the dialoguehas turned




into monologue as deinstitutionalizationand centristic

commandhave grown apace.
Each presidentin recenttimes has begun office with
the supposition that the governmenthas no organic
past. At each turn, the wheel is to be reinventedanew.
At their core, argumentsfor furtheringthe process of
politicization and centristic command also conclude
that leadership is equivalent to the introduction of
novelty and that institutionalizationis an obstacle to
Since politicians are constitutionallyempoweredto
direct government,there can be no argumentthat the
administrativeapparatus, other things being equal,
must be responsiveto the politicalleadership.The question is what that responsivenessmay mean and what,
therefore,is the responsibilityof the seniorcivil servant.
We quote here from our earlierstudiesthe reactionof a
Germancivil servantto this problem:

their own appointees. Even more, it is a matter that

presidentialadministrationsmust define in the context
of other institutions that the American system constitutes as authoritativeprincipals.Thus, it turns out
that the real issue often is not politics versus neutral
competence but clarifying the principals (and their
underlyingprinciples)in the principal-agentrelationship. Politicizationand centralizationare appropriate
presidentialresponsesin efforts to define the terms of
the relationship-to a degree.Beyondthat unspecifiable
point, however, strategies for achieving presidential
responsivenessturn into tactics for exclusivepresidential rule. Efforts to achievethat level of aggrandizement
areruinousfor governancein the Americansystem;that
is, they are collectively irrational. They also are
ultimatelyruinous for presidentswhose political wellbeing probablyis essentialfor effectivegovernanceand
are thus likely to be individuallyirrationalas well.

We are not hereto receiveorders,mentallyto click our heels, and to

say "Jawohl!"-that's not why we are here. On the contrary,if
(seniorcivil servants)have a differentconception(of the problem)and they shouldalwayshavea politicalconception-they mustunder
certaincircumstancesuse their conceptionin conjunctionwith their
expertiseand simplysay, "But I wouldproposethusand suchfor this
reason."And if the ministersays, "No, politicallywe can'tdo thaton
account of these reasons," then all right, it alreadywill be done as
proposed(by the minister).It mustbe this way, becausethe ministeris
the responsibleofficial, who must have the last word. That can't be

At the basis of the contention that furthering politicization of the bureaucracy is in

the collective interest is the belief that
presidential leadership is essential and
whatever enhances it is a good.

Even though senior career executivesin the United

States are more likely to be talking to assistantsecretaries instead of the ministerial equivalents of their
departments,it is not difficultto imaginediscussionsof
the sort exemplifiedin the quote taking place much of
the time.
Althougha good many claimshave been made about
the recalcitranceof careercivil servantsto follow the
policy and programcoursethat a presidentialadministration is embarkingon, little evidence supportsthese
assertions when effective administrativeleadershipis
brought to bear. Good management, as reflected in
open channelsof communication,willingnessto listen
to advice, clear articulation of goals, and mutual
respect, in fact, may also constitute good politics for
departmentsecretariesor their assistantsecretaries.No
evidenceshows that good managementis incompatible
with effective politics unlessthe impositionof stringent
commandproceduresis regardedas an integralpartof a
presidentialadministration'spolitical style. The antibureaucraticstyles of recent administrationssuggest
that this symboliccomponenthas becomeat least as importantas achievingresults.
Responsivecompetencefrom the executiveapparatus
is a legitimaterequestof presidentsup to the limits we
have described.No one seriouslywould arguethat the
administrativemandarinateshould be unaccountable.
So, the issue is what can, and should, presidentstry to
control. That, it turns out, is a matterthat presidential
administrationsoften must settle internally amongst


The key issue, therefore,is not whethersome degree

of politicizationis necessaryto promoteresponsiveness,
but ratherhow much. The issue is not whetherresponsivenessshouldbe promoted,but ratherhow reflexively
and to whom. The model proposedfor more presidential aggrandizement,ironically,is a prescriptionto rob
governmentof its capabilityfor realitytesting, and it is
withoutdoubt a model for demoralizationof the career

What Is in a President'sInterests?
The argument that presidential command of the
bureaucracyneedsto be furtheredis rooted in the value
ascribedto presidentialleadershipand in the view that
presidentialgoals and directionsare overriding.In this
view, the bureaucracyneeds to be mobilized in accordancewith these goals and directions.At the basisof
the contention that furthering politicization of the
bureaucracyis in the collectiveinterestis the belief that
presidential leadership is essential and whatever
enhancesit is a good.
While we believe that Terry Moe's analysis also is
sympatheticto this generalview, his more fundamental
argumentis that presidentsineluctablyare driven to
politicizationand centralizationbecauseof the relationship betweenstructuresand incentivesin the American
As Moe asserts:
In an ideal world, presidentsmight pursuea varietyof institutional
reformsin rightingthe imbalancebetweenexpectationsand capacity.
In the real world, they readilyembracepoliticizationand centraliza-


tion because they have no attractivealternatives.The causes are

systemic-they are rooted in the way the larger institutionalized
systemis put together.'3

Two points are necessaryto address because they

representimportantambiguitiesin any analysis of the
subject of presidentialprerogativesand the use of the
executive. The first is what it is that constitutes
politicizationand centralization.The secondis the need
to distinguishbetweenthe apparentincentivesa president has (or more properlyis inclined to see) and his
The first point is especiallydifficult. It is impossible,
we agree, to deny the need for politics or for political
leadershipof the administrativeapparatus.However,
the reverseargument,that which implicitlydenies the
need for deliberation,skepticism,and continuity, has
become more frequent.What makesthis issue so complex is not the readily agreed upon notion that the
bureaucracyrequires political leadership and supervision, but the problem of defining the legitimate
thresholdsof this. At what point, for example, should
an issue be politicizedin decision making?

Precisely because the president and presidential appointees in the executive have
such short time horizons, the norms of
cooperation are difficult to develop,
especially once noncooperative norms of
behavior have taken hold.
Rather than the broad argument as to whether
politicizationand centralizationare goods or bads, we
need to specify the mechanismsand also the political
conditions under which these operate. Some mechanisms are legitimate;others are not. Some may be wise;
others are not.
With regardto the secondpoint-that of presidential
incentivesand interests-we distinguishdifferent conceptions of "interest." The discipline of economics
tendsto definea person'sinterestsby whatone is willing
to pay for. Interesthas an operativemeaning. Therefore, by this logic, how presidentsbehavein a situation
expressestheir interest.When they behaveso as to aggrandizepower,that expressestheirinterestand reflects
the structureof incentivesaroundthem. But presidents,
like consumers,make choices with uncertaininformation. Put in front of a candycounter,a child is likelyto
make dietary decisions inconsistentwith his interests.
Whenpresidentscome to office withouthavingbeenexposedto careerofficials, but often only to horrorstories
told about them, they too may make decisions inconsistent with their interests.
The fact is that presidentscan get into verydeeptrouble when they do end-runs around the bureaucracy,
when commandreplacesdeliberation,and when White
House centrismbringsforth the illusion of centralcontrol. Nixon's fall from power was paved by the Watergate break-in,but it had as much to do with abuses of


the executiveas anythingelse. Even had Watergatenot

occurred,but with Congressremainingin the hands of
the Democraticopposition,it is hardto imaginethat the
congressionalhand would have been stayed for long.
The revelationsof 1986-87involvingthe White HouseNSC operationof armsshipmentsto Iranand laundered
funds to the Nicaraguancontrasalso threatensto erode
fatallythe politicalstandingand the policy credibilityof
the Reagan presidency. Operating through the back
door and around the institutionalizedapparatus of
governmentcan lead to decisionsand illegalitiesthat are
truly presidency-threatening.
It is hard to imaginethat
this is in a president'sinterests.
One of the major functions, in short, of the permanent apparatusis to serve presidentsby helping them
avoid stupidmistakesthat threatentheirpoliticalviability. The urgeto commandand to centralizeoften fails to
recognizethat politicalimpulsesshould be subjectedto
tests of sobriety.Thoughthereare a good manyreasons
to argue on behalf of the basic idea of "neutralcompetence" and againstthe politicizationof all executive
organizations,the most fundamentalone that a president ought to consideris the avoidanceof errorand illegalitythat have wrackedrecentpresidencies.

Conclusion:Monopolyand Competition
in AmericanGovernment
As we readthe insightfuland provocativeanalysesof
RichardNathanand TerryMoe aboutthe need for more
presidentialism(or, in Moe's case especially,the needs
of presidentsthemselves),we are struckby how similar
their and our descriptionsof the Americansystemare.
We see, as they do, a systemof intensecompetitionfor
resourcesin the strugglesto define publicpolicy and to
jockey for political advantage.In broad contours, the
systemlooks to us (two centuriesremoved)as Madison
hoped it would. The competitive struggle leaves no
single institutional actor with sufficient resourcesto
fully dominate the system in the absence of extensive
and deep consensus.
The analysesof Nathanand Moe, while imbuedwith
some novel twists, fit broadly into a long line of
presidentialistliteraturethat urges reformto make the
systemmorecompliantwith presidentialobjectives.The
difference, as Moe indicates, is that most of that
literature is organized around nonexecutive reforms
whose prospectsare implausible.The only significant
tools available, according to this logic, are executive
ones-politicizing the bureaucracyand centralizingexecutive command. In essence, presidentsdo what they
have to do with what they have available.But the spirit
of presidentialismis the motivatingideal. In the end, it
is the president on whom falls the responsibilityof

That being the case, presidentsneed, in this line of

analysis,to maximizetheir advantagesin a systemthat
endows them with too few. Maximizing advantage
throughthe executive,in Moe's view, is a normthat has
evolved because presidentsincreasinglyhave found it




essentialas a means of accomplishingtheir goals. The

trouble with this norm, among other things, is that it
tends to induceretaliatorybehavior.WhenU.S. Office
of Management and Budget (OMB) or presidential
emissaries decide to rewrite regulations to fit their,
rather than statutory, definitions of policy, Congress
will retaliate when it has the political will. Because
presidentshavethe advantageof initiativein thesesituations, however, they may see little to lose in pressing
that advantage.But retaliatorybehavior-and with it, a
loss of credibility-has a good chance of being provoked.
In the short run, the system,as Moe argues,provides
incentives for maximizing advantage, and since the
players,especiallythe presidentialones, are short-term
actors, it is understandablethat these incentivesseem
compelling. Norms have evolved in the White House,
particularlyamong Republicanpresidents,to politicize
and centralizethe executiveapparatusin especiallyexuberantfashion. But other normscan evolve as well if,
in the long run, ceaselesspoliticizationand centrismare
seen as having disadvantages.
Through his experiments, Robert Axelrod draws
some interestinglessons about how norms of cooperation evolve. In Axelrod'smodel, whichhe calls TIT for
TAT, time and the continuityof relationshipare important elements.'4Negative sanctions must be timely so
that they can be linked clearly to a player's move to
defect. Thus, we can infer that usingthe executivein illegal ways should be met more swiftly than not with
congressional or judicial retribution. A larger time
horizon is necessary,however,to ensurethat a benefit
to improving a continuous relationshipis perceived.
Whenthe marginalcost to defectis low, stemmingfrom
a failure to retaliatein a timely way, and, above all,
from a belief that a relationshipis noncontinuous,it is
difficult for norms of reciprocityand cooperationto
Of course, the extent to which Congress or the
judiciarywill reactwill dependlargelyon the prevailing
politicalclimate,and to the extentthat thereis reaction,
it likely means that senior career officials will be
squeezedfrom all sides. That is not likelyto be a condition that enhanceseitherthe statusor the role of career
officials or the qualityof governance.And the slowness
of reactionundermost circumstancesmeansthat presidents often learnthe necessarylessonslate, perhapstoo

The incentives toward reciprocity need to be

strengthened.If presidentsare quickly and forcefully
remindedabout what they cannot as well as about what
they can achieveby efforts to monopolizeinstitutional
power through command, perhaps, then, they will be
more inclined to seek other means for influencing a
governmentthat they only partiallyhead and whichhas
an executiveapparatusthat is not undertheir exclusive
control. Respectfor that principlemay turn out to afford presidentstle best opportunityto achieve their
goals without recurrentbacklash. In a system such as
ours, it is vital to develop norms of cooperative
behavior.That, of course,is a differentmodel of how a
systemstructuredaroundcompetitionmight work.
It is hard, however, to be optimistic about this.
Precisely because the president and presidentialappointeesin the executivehave such short time horizons,
the norms of cooperation are difficult to develop,
especiallyonce noncooperativenormsof behaviorhave
taken hold.
This is the crux of a crucialcurrentdilemmafacing
the Americanpresidencyas an institution.If presidents
follow their short-term interests, they are likely to
stimulatemoreand morerestrictivecongressionalbonds
on their behavior, therebygiving presidentsincentives
to engage in the types of behaviorexemplifiedby the
Iran-ContraAffair. Yet each individual president is
likely to put his short-terminterestsabove the institution's interests.As in many other aspects of American
politics, Congressis key here. It will ultimatelydetermine the kind of presidencywe get. It must act expeditiouslywhen presidentsarrogatefor their exclusive
use constitutionally shared authority. Otherwise,
presidentswill take as theirswhat Congressby its inaction bestows.

JoelD. Aberbach
is Professorof PoliticalScienceand
Program Director in the Institute for Social Science
Researchat the Universityof California,Los Angeles.
His currentresearchproject(with Bert A. Rockman)is
a study of the changingfederalexecutive.
BertA. Rockman
is Professorof PoliticalScienceand
ResearchProfessor,UniversityCenterfor International
Studiesat the Universityof Pittsburgh.He is currently
involved(withJoel D. Aberbach)in a studyof executive
and politicalchange in Washington.

The authorsaregratefulto Paul Quirk,TerryMoe, MarkPetracca,
and MichaelReaganfor commentson an earlier,moreextendedversion of this paper.
1. RichardP. Nathan, "InstitutionalChangeUnder Reagan," in
John L. Palmer, ed., Perspectives On The Reagan Years
(Washington:The UrbanInstitutePress, 1986),p. 121.
2. Richard Rose, "GovernmentAgainst Sub-Governments:A
EuropeanPerspectiveon Washington,"in RichardRose and


Ezra N. Suleiman, eds., Presidents and Prime Ministers

(Washington: American Enterprise Institute, 1980), pp. 284-347.
3. Nathan, "Institutional Change," p. 128.
4. Nathan, "Institutional Change," pp. 133, 132.
5. Terry Moe, "The Politicized Presidency," in John E. Chubb
and Paul E. Peterson, eds., The New Direction in American
Politics (Washington: The Brookings Institution, 1985), pp.



6. RichardRose, TheProblemof Party Government(New York:

The Free Press, 1974).
7. TheodoreJ. Lowi, TheEnd of Liberalism:TheSecondRepublic
of the UnitedStates, 2d. ed. (New York:W. W. Norton, 1979).
8. BertA. Rockman,"TheModernPresidencyandTheoriesof Accountability:Old Wine and Old Bottles," Congressand the
Presidency,vol. 13 (Autumn1986),p. 138.
9. RichardE. Neustadt, "Politiciansand Bureaucrats,"in David
B. Truman, ed., The Congress and America's Future
(EnglewoodCliffs, NJ: Prentice-Hall,1965),pp. 102-120.
10. Speakingof both the WhiteHouseand CapitolHill, in regardto


directionof the bureaucracy,Neustadtcomments(from present

perspectives,ironically)that "at both ends of the Avenue, to
urgeawarenessof joint stakesand commonrisksis not perhaps
to ask too much of our establishedsystem."Neustadt,p. 120.
Joel D. Aberbach,RobertD. Putnam, and Bert A. Rockman,
Bureaucratsand Politicians in WesternDemocracies(Cambridge:HarvardUniversityPress, 1981),especiallyChapter8.
Aberbachet al., p. 249.
Moe, p. 246.
Robert Axelrod, The Evolution of Cooperation(New York:
Basic Books, 1984).

Frontiersof PublicService
49th NationalConferenceApril16-20, 1988

SPA's 49th National Conference
will be a stimulating and
professional event. All those who
believe in the value and importance of
government and the dignity and worth
of public service should be there.
If you are not a member of the American
Society for Public Administration and
would like to receive informationregarding
"Frontiers of Public Service," ASPA's
49th National Conference in Portland,
Oregon, please complete this form. We will
be happy to rush registrationmaterials off
to you.




This form should be mailed to:

Department of Conferences

and ProfessionalDevelopment
1120 G Street, NW
Suite 500

DC 20005