a model of political leadership - control...

A model of political leadershipKevin Morrell and Jean Hartley

A B S T R AC T In this article we develop a model of political leadership. In doing so,

we analyse the challenges facing political leaders in local government

in England and Wales. We use this analysis as a basis for broader theo-

rizing: about leadership at other levels of government, and in other

countries. The scope for applying extant accounts of leadership in

these domains can be enhanced by considering the relational

complexities that characterize the environment within which political

leaders act; by doing so we offer an agenda for research. We describe

the context for political leaders in terms of figurational sociology,

where figurations denote interdependent networks of social relations.

These take shape in different arenas of action, and are partly influ-

enced by the different roles that political leaders undertake. These

figurations are also constituted differently given the diversity inherent

in the context for enacting political leadership. We propose a concep-

tual model that serves both as a heuristic framework to organize

conceptualization of this understudied area, and to orient future

research.

K E Y WO R D S figuration � government � leadership � model � political

Introduction

Although there is a large literature on leadership, much of this is acontex-tual, framed in terms of the individual, and atomistic. In Fairhurst’s terms(2001: 383), the ‘dominant views of leadership have been shaped by atraditional psychological view of the world where, in a figure-ground

4 8 3

Human Relations

DOI: 10.1177/0018726706065371

Volume 59(4): 483–504

Copyright © 2006

The Tavistock Institute ®

SAGE Publications

London, Thousand Oaks CA,

New Delhi

www.sagepublications.com

arrangement, the individual is figure, the system is background’. This dualismprecludes systemic perspectives (Senge, 1990), and accounts that acknowl-edge the interplay between structural and agentic explanations of socialphenomena (Giddens, 1984). As well as potentially oversimplifying associ-ated phenomena, such as charisma, or effectiveness, the ‘traditional psycho-logical view’ privileges research from nomothetic paradigms, at the expenseof interpretivist perspectives such as ethnography and social construction-ism. This can bias research towards causal and functionalist accounts, at theexpense of more detailed exploration of meaning, and intentionality(Ghoshal, 2005). It is important to address these limitations, since the wayin which social phenomena are labelled and explained influences themthrough the Pygmalion effect (Ferraro et al., 2005). Here we offer an accountto complement rather than usurp established accounts of leadership. Ourapproach has broad implications although the aim is to develop research inan understudied area, by focusing on political leaders: those who hold formalpolitical authority, acquired through election in a democratic society.

In the organizational behaviour literature, empirical research has oper-ationalized ‘political skill’ as a discrete subset of behaviours that can enhanceeffectiveness, or explain career success (Moss, 2005). ‘Political power’ hasbeen used narrowly to refer to the ways in which decisions are controlled,and resources are deployed and distributed in organizations (Yukl, 1989). Ineveryday discourse, ‘politics’ is both a signifier and explanation for a rangeof unfavourable workplace outcomes; ‘politicking’ describes occasionallyunsavoury activities such as lobbying or ‘dirty tricks’. In both senses, thesepejorative connotations of politics can colour perceptions, and muddyanalysis of political leadership. We argue here that political leadership is asubstantive area of activity that is understudied in the contemporary leader-ship literature, and as such it warrants attention.

Political leadership

Here we define political leaders as: i) democratically elected ii) representativeswho iii) are vulnerable to deselection, and iv) operate within, as well as influ-ence a constitutional and legal framework. Their source of authority is v) amandate: ‘permission to govern according to declared policies, regarded asofficially granted by an electorate . . . upon the decisive outcome of anelection’ (Chambers dictionary, 1993). Membership of the electorate is vi) setout in law, and broader than organizational or union forms of membership,since it extends to all citizens with voting rights, in a defined constituency.

This summarizes the basis for political leaders’ claims to authority. Itillustrates how they differ from other leaders such as: chief executives,

Human Relations 59(4)4 8 4

managers from the private, public and voluntary sectors and those in themilitary. The definition is brief and contains a number of terms that couldwarrant further discussion, or are understood relationally (electorate, citizen,constituency) but we propose the above criteria as necessary conditions thatare broad enough to include many different types of political leaders. Itcovers what we call ‘formal’ political leaders, though not ‘informal’ politi-cal leaders, such as those who head special interest groups, or civil rightsactivists, or trade unionists (Azzam & Riggio, 2003). Nor does it coverdespots (Kets de Vries, 2005).

Since political leaders are elected rather than appointed, and act asrepresentatives, they require consent from those whom they govern andserve. They have a duty to serve all their constituents and protect the inter-ests of future generations, rather than simply those who supported them.This should include the elderly and disadvantaged groups, as well as thosewho do not have the power to vote, such as children. These are typical differ-ences in comparison with non-political leaders, but political leaders alsooperate under different structures of accountability and scrutiny. In addition,they have formal legal responsibility for a broad range of issues: health, lawenforcement, taxation, education, legislation and the economic sphere. Thenetworks within which they act have regularities but are also fluid. Politicalleaders gain authority through the ballot box initially but their authority ispotentially subject to challenge on a daily basis, from: their political party(most operate within a party structure), opposition politicians, the media,their constituents, and other bodies (e.g. charities, lobby groups, businessconfederations).

One characteristic of the challenges facing political leaders is thatactions and decisions may require the mobilization of different groups inorder to build consent. In this sense, the issue for politicians is to gain someconsensus across the entire domain of a problem (Thompson, 1967). Afurther source of complexity is that political leaders are directly responsiblefor the provision of public services (e.g. street lighting), and also have a regu-latory and enforcement role (e.g. in collecting taxes). The ubiquitouslanguage of customer relations breaks down here since many ‘customers’ ofregulatory services are unwilling ones (Pollitt, 2003). People’s expectationsof what leaders can provide may differ widely from what is actually possiblegiven legal, logistical and practical constraints (Boin & Hart, 2003). Thesedifferent challenges show how the relationship between leaders, stakeholdergroups and the electorate is complex and interdependent.

There are some areas of overlap and similarity here with non-politicalleaders, such as senior chief executives (CEs), and leaders in publicadministration (Cook, 1998). CEs may feel they, and their organization, havea moral duty to their stakeholders, a term that can be interpreted broadly

Morrell & Hartley A model of political leadership 4 8 5

(Kelly et al., 2000). Accordingly, they may believe they should act in theinterests of all stakeholders, even those who are opposed to their appointmentand strategies. Governance structures within an organization may be demo-cratic in the sense that those in authority are accountable to their superiors,who are in turn accountable to collective, representative bodies (Ackoff,1994). These arrangements could be extended to include other stakeholders,bringing the model of a democratic corporation closer to the literal sense ofdemocracy, rule by the people (Skelcher & Mathur, 2004). CEs may seek toinfluence policy, to mobilize coalitions and consolidate support, for instanceprior to a shareholders’ meeting. Writers such as Charles Handy have arguedthat the boundaries between public and private sector organizations areblurring, an issue that has been explored in greater depth with reference tothe public sector generally (Robbins, 1993), and local government in particu-lar (Benington, 2000). Hartley (2000) argues that politically led organizationscan provide useful insights into generic processes in organizational behaviour.Similarly, Brunsson (1985: 116) suggests political organizations, ‘exposefundamental problems connected with rationality and action and can teachus a great deal about fundamental problems and solutions in organizations’.

Despite areas of overlap and similarity, we submit that the definitionabove describes a particular kind of activity, one broad enough to referencea large number of leaders for whom leadership is sufficiently coherent anddistinctive as to be considered a phenomenon in its own right. The definitionorients research away from an exclusive focus on behaviours because itconsiders the formal basis for authority, and the context within which leader-ship is enacted. We do not question that the same behaviours may manifestthemselves in political and non-political leaders, however, given the particu-lar structures within which political leaders operate, attention to context isalso important (Hoggett, 2006; Leach et al., 2005). Though leaders fromother sectors face complex challenges, they do not have a similar mandate,and this is not addressed in the contemporary literature. Here we address thisgap by offering a model for political leadership that is sensitive to structuralconstraints, embedded in the context, and constituted in social interaction(Collinson, 2005). Prior to exploring the suitability of existing theories ofleadership, we outline the particular contextual complexities facing politicalleaders.

Political leadership in situ

In this section, we propose and discuss a summative framework for politicalleadership. This is developed from consideration of a particular context: localgovernment in the UK. To maintain clarity, subsequent exposition draws on


Morrell &

Hartley

A m

od

el o

f po

litical le

ad

ersh

ip4

87

Diversity Index Ranged profile of leader behaviours

Demography Scale of change, social diversity No special requirements … Sensitivity, cultural (population profile, social change) awareness, ability to engage diverse groups

Community Level of social cohesion Foster community initiatives … Encourage wider (consumerism/expectations) levels of participation

Economy Levels of wealth, educational Engage poorer parts of community, lobby for (level of wealth, employment) attainment, employment funding … Foster growth, satisfy electorate

Government Relations with other layers of Consolidate, build on regional/national identity … (local, regional/national agenda) government Negotiate greater regional/national autonomy

Partnerships Extent of partnership Providing services, fostering partnerships … (providing > influencing role) arrangements Fiduciary competence, collaborative skills

Control or opposition Majority/NOC/minority Leading/influencing & persuading … Role model (political stability, influence, culture) Entrenched/progressive for change/facilitative – empowering, persuading

Political allegiance Ties and lo yalty to party or other Relative independence from party … Close ties(to party, to government) political bodies

Legitimacy Levels of representativeness and Build on democratic legitimacy … Increase (accountability, open government) turnout visibility and links with voters

Modernization Perceived effectiveness at Consolidate strengths … Implement reforms to (reform agenda and mandate) innovation and improvement enhance improvement

Technology Degree to which technology is Enable and promote learning and access … Drive (eGovernment) integrated, used and understood forward innovative technologies

Size Population served, number of Managing bureaucracy, centralization … (population, size of council) officers, civil servants Leveraging limited resources

Rolea

Executive

Scrutiny

Regulatory

Advocacy

Responsibilities

Key decision-makers,heart of political body

Investigate broad policy, challenge andexamine decisions,review

Weighing anddetermining applications andappeals; directregulation (discretion)

Responsibility of allpoliticians to theirelectorate

KSAs valuesb

Delegation, informationprocessing, persuading,influencing, ‘Gets thingsdone’

Meeting skills, research,influencingCritical questioning

Representation,research, judgementAnalytical skills, care,attention

Listening, advising,facilitating, researchInterested in people,service ethos

a Information on roles, responsibilities, profile and context is drawn from Stewart (2003), roles found at local, regional, national and international level.b KSAs = Knowledge, skills, attitudes.

Figure 1 Dimensions of contextual diversity for political leaders

this context, though as we argue, this summary is potentially applicable toall forms of political leadership as earlier defined (see Figure 1). Theframework represents the inherent contextual diversities faced by politicalleaders. It illustrates these across several domains, which influence howleadership is constructed and enacted.

Column 1 shows four generic roles which political leaders undertake:executive, scrutiny, regulatory and advocacy (Stewart, 2003). These areformal descriptions of leadership roles for local politicians in England andWales, but can reference activities common to political leaders in othercontexts, at other levels of government, and in other countries. For each,there is a range of responsibilities, some of which may be part of the role,while others may be self-chosen (column 2). From the responsibilities can bederived a role profile, where a range of knowledge, skills and attitudes (KSAs)are desirable (column 3). Importantly too, political leaders should share someof the values that pertain to living in a democratic society, for example,accepting diversity, and seeking to take into account the interests of allconstituents not just those who share one’s political allegiances. They shouldalso be people of integrity and probity since they are responsible for admin-istering public monies and making decisions that have an impact oneconomic and social well-being.

Alongside these generic profiles, we list several dimensions of thecontext for their leadership, the ‘contextual filter’ (as bracketed). These willvary from place to place, and differences will influence their constituents’,and leader, priorities. Column 4, ‘diversity’, lists each dimension. Column 5,‘index’, operationalizes the ways in which these may differ. We suggest eachdimension can be operationalized with ordinal or scalar measures. Column6 briefly hypothesizes different behaviours that are likely to prove more orless effective in the light of the particular context. This account does notimply a mapping of the context for each political leader; instead, it illustratesdimensions along which the context is in flux. This offers greater scope thanmodels that are static or exclusively focused on content, or that fail toacknowledge the importance of structural elements.

To develop Figure 1, below we describe different contextual featuresthat shape the domains within and across whose boundaries politiciansoperate. To provide clarity and sufficient detail, this is illustrated using thepolicy and academic literatures dealing with local government (Hartley &Benington, 1998; Wilson & Leach, 2002). This preserves a consistent levelof analysis in the sections that offer in-depth description of contextualcomplexity. However, we contend that the logic informing this account ofpolitical leadership, applies at regional, national and supranational levels,and that Figure 1 stands as a visual summary that is transferable to these


contexts. Our account describes three arenas of diversity, respectively: the‘place’, whether local, regional or national; the ‘authorizing environment’(Moore, 1995), and the ‘organizational environment’, for example, the localauthority, a regional assembly, or national and supranational parliaments.Within each category are a number of sub-categories that influence thediversity of the leadership task.

The ‘place’

Some local governments in the United Kingdom (UK) have a relatively stablepopulation where the pace of demographic change is low, for example,certain rural settings with limited inward migration. Other councils face chal-lenges of coping with, for example, national growth plans, economicmigrants or asylum seekers. Changes may put pressure on the provision oflocal public services and the political choices this requires. Political leadersmay have to be able to represent and articulate the views of a diverse popu-lation. The level of social cohesion also varies immensely. In some areas,where there is a strong sense of community, political leaders may act in afacilitative role, consolidating existing networks; in other areas there will bea challenge to reach out and create a sense of community. The politicalleadership roles and challenges are likely to be different in these differentcircumstances (Leach et al., 2005). Finally, the economy affects challengesfor political leaders. For example, with high unemployment and socialdeprivation there may be associated problems of low educational attainmentand crime (Lupton, 2003) and a key role for political leaders may be toattract extra resources. Deprivation may also have a depressive impact onaspirations. In areas where there is high employment and prosperity, the localcommunity may create greater demand for more complex service provision.

The ‘authorizing environment’

In the UK, local government is highly dependent on central government forlegislative and financial freedom. This contrasts with the USA, for example,where authority essentially derives from local government upwards to stateand federal bodies. In the UK, there is variation between councils in thedegree to which they have achieved some (limited) degree of autonomy fromcentral government. Shifts in emphasis in the role of local government fromservice provider to collaborative partner in service provision have beenpronounced in the last decade (Leach & Wilson, 2002). Some authorities arewell practiced in developing suitable partnerships to enact these changes(Geddes & Benington, 2001). In others, political leaders are building these


relationships from scratch, which requires a different set of skills andleadership processes. The distribution of power within a local authority mayinfluence the repertoire of behaviours politicians can deploy. For example aminority group opposition faces different challenges to a majority partyadministration. This may affect the relative emphasis leaders place on coreskills, for example, influencing, facilitating, and persuading. Allegiancediffers across councils according to how integrated the controlling group iswith their political party’s national structure (Leach & Wilson, 2002). Thiscan influence the degree of freedom with which a political leader can operate.Finally, different factors affect the legitimacy of leadership. First, electoralturnout varies across councils and can adversely affect the legitimacy withwhich politicians govern (Rallings et al., 1996). Second, where councillors areunrepresentative (in terms of demographic characteristics) this may reduceperceptions of legitimacy (Employers’ Organization & I&DeA, 2001). Third,political leaders act in different roles: enforcing and influencing regulation,being advocates, and having responsibility for service provision. This canmake it difficult to articulate the relationship between leaders and thecommunities they govern (Ranson & Stewart, 1994). There may be diverseinterpretations of local politicians’ accountability (Pollitt, 2003). Each facetof legitimacy illustrates how leadership is socially constituted.

The ‘organizational environment’

The challenge of responding to pressures for reforms is met differently acrossdifferent politically led organizations, and performance across these alsovaries. Authorities are likely to be in a position of consolidating on existingstrengths, or working hard to reform. Irrespective of the legitimacy ofperformance ratings (Boyne & Enticott, 2004), these influence stakeholderperceptions and affect the priority political leaders give to internal organiz-ational improvement (Leach et al., 2005). Local authorities also vary in theirability to deploy technologies. These can enhance local democracy andimprove the ability of authorities to link with other agencies (Martin, 1997).Though identified as an explanatory variable for a range of phenomena(Child, 1984; Pugh, 1981), few studies explicitly consider the relationshipbetween organizational size and leadership. However, there is reason toexpect that the size of a local authority and its population are likely to influ-ence the behaviours of political leaders (Boyne, 2003). For example, thelargest local authority in England serves over a million people, with morethan 100 elected politicians. Smaller local authorities may serve a populationof under 30,000, with fewer than 20 politicians. This variety has implicationsfor the structure of the organization, but also for governance arrangements,


and the management of the authority. Having discussed the particularcharacteristics of, and contexts for, political leadership, the following sectionbriefly reviews the current leadership literature to assess its suitability forstudying this phenomenon.

The literature on leadership

Grint’s The arts of leadership (2000: 2–3) discusses four kinds of leadershiptheory: trait approaches, contingency approaches, situational approachesand constitutive approaches. He categorizes these according to whether theyemphasize the individual or the context as essential. Trait approaches are‘essentialist’ in terms of the leader but ‘non-essentialist’ in terms of context:‘a leader is a leader under any circumstances’. Contingency approaches areessentialist in terms of the leader and the context: ‘both the essence of theindividual and the context are knowable and critical’. Situational approachesare essentialist in terms of the context, but not in terms of the leader: ‘certaincontexts demand certain kinds of leadership – so we do need to be very clearabout where we are’. The fourth kind, constitutive approaches, are non-essentialist in terms of the leader and the context: the meaning of contextand leader are both contested, ‘leadership must still be perceived as “appro-priate’’, but what that means is an interpretive issue’. We use this frameworkto examine briefly the academic literature on leadership, with a view toassessing its suitability for understanding political leadership.

Trait approaches

Much leadership research views leaders as individuals with a particular setof traits or abilities, or sees leadership as the exercise of particularly effec-tive sets of behaviours. Leadership dimensions have been developed forexample for health service managers (Alimo-Metcalfe & Alban-Metcalfe,2001) and trait-based approaches have informed the study of union leaders(Metochi, 2002). Similarly, recent theories hypothesize linkages between typeof leader and a dependent variable(s). For example, transformational leader-ship traits have been linked with: knowledge creation, organizationalperformance, follower self-concordance, work alienation, creativity, andhigher order motives (see Vera & Crossan, 2004 for a review). Other studiesexamine charismatic (Javidan & Waldman, 2003) and transcendental leader-ship (Sanders et al., 2003) in a similar fashion. This approach exhibitsparadigmatic consensus (Pfeffer, 1993); that is, given terms (transformationaland transactional leadership) are commonly understood, and there is a


methodological consensus – using inventories to assess the type of leader-ship, and a range of measures to assess the dependent variables (Barker,2001). One advantage of having paradigmatic consensus is that there is lessdebate over first assumptions. However, this can be a limitation if it leads tothe monolithic representation of phenomena, or leaves little room fordiversity in approaches (Zald, 1997). One reason that trait accounts arelimited is that they underplay the importance of leaders’ interdependence.For political leaders, effectiveness and identity is not merely a function ofcharacter, but socially constituted in a rich, dynamic context, and subject tostructural constraints and opportunities.

Contingency approaches

The most influential contingency model is by Fiedler (1967), which empha-sizes leadership style in combination with the relevant features of a givensituation. This suggests that performance is a function of leaders’ style andsome key features of the context (leader–follower relations, task structure,position power). A contemporary example offered by Liu et al. (2003: 127)suggests leadership styles should be matched to the modes of working ofemployees, as determined by the ‘objectives and psychological obligationsunderlying different groups of employees’. Other contingency approachesconsider top management teams (upper echelons theory), or shared leader-ship, where the characteristics of an elite group are viewed as appropriateto a particular context (Ensley et al., 2003; Papadakis & Barwise, 2002).Contingency approaches have an advantage over trait accounts in that theyacknowledge the role of context. However, they also imply a mechanisticrelationship between style and context, as suggested by notions of‘matching’ and ‘fit’. This is an instance of a dualist framing of leadership,between ‘the individual and the system’ (Fairhurst, 2001). The difficulty ofapplying this to politicians is that the context is constantly changing andcan be different depending on the nature of particular challenges, or thedifferent actors with whom they interact. Political leaders’ networks are notstatic, as is implied by the idea of a prevailing mode of working, insteadthey are highly fluid. Changes in these, or the wider context, influence thecontingencies of leadership.

Situational approaches

There are two prominent theories of situational leadership. Vroom andYetton (1973) identify five leader decision-making styles, and Hersey andBlanchard (1988) suggest leaders adapt styles to suit the readiness of their


followers to perform. In both models, the context determines which style isappropriate. Situational approaches, like contingency approaches, acknowl-edge the importance of context and thus offer advantages over trait accounts.However, they overlook the ways in which leader and context may be inter-dependent. This is a limitation because political leaders are concerned withdeveloping far-reaching policies that govern the ‘authorizing environment’(Moore, 1995) within which organizations and institutions operate. Thismakes it harder to treat ‘context’ for political leaders as a given (Leach etal., 2005). Managerial leaders typically promote the interests of an organiz-ation, or less typically, a coalition of organizations with shared interests.Though managerial leaders have a moral duty to consider what may beloosely called ‘stakeholders’, political leaders bear the social expectation thatthey will consider the impacts of their decisions on every group in theirconstituency, and sometimes beyond. These groups may have competingvalues and goals, so the challenge is to address difficult problems where thereis conflict as to what the criteria informing decisions should be, and lessconsensus about desired outcomes (Burns, 1978; Heifetz, 1994). There is alsolikely to be disagreement over how outcomes are measured (Behn, 2003).This in turn implies interdependence between leaders and other actors in anetwork, which an exclusive focus on trait, situational or contingencyapproaches does not provide.

Constitutive approaches

The constitutive approach contests the objectivity of the terms ‘leader’ and‘context’, emphasizing the role of interpretation and the way in which theseterms are interrelated (Denis et al., 2001). Though there is less literature onthis approach, the logic underlying it can partly be seen in leader memberexchange (LMX) theory (Dansereau et al., 1975). This holds that leaders usedifferent leadership styles with different groups of subordinates (Sparrow &Liden, 1997), and where assignment to a favoured ‘in-group’ results infavourable, informal exchanges between leaders and members, whereas ‘out-group’ interactions are more task-oriented and formal. Following theseanalyses, the leader and the context (in terms of work group) are definedrelationally (Leach et al., 2005). One of the potential benefits of a constitu-tive approach is that it allows space to explore the ways in which leadershipis socially constructed (Grint, 2000, 2005; Heifetz, 1994). This opens oppor-tunities for alternative methodological paradigms, such as ethnography, andmethods such as discourse or narrative analyses (Putnam & Fairhurst, 2001).A limitation is that exclusive focus on constitutive accounts makes it harderto pursue prevailing nomothetic paradigms based on quantitative methods.


Developing the study of political leadership

The brief review above suggests there are shortcomings in each approachidentified by Grint’s study of contemporary leadership research. Though weaccept these may adversely affect the study of other forms of leadership, thereis good reason to address these shortcomings as they concern study of politi-cal leadership in particular. Political leadership is a relatively understudied,though discrete and substantive area of interest and in this sense there is morescope to influence future research than in areas where there is a well-estab-lished body of literature (McKinley et al., 1999). In addition, for theoreticalreasons, it is interesting to explore a phenomenon where: i) structural simi-larities (as outlined in the earlier definition) persist across different levels, anddifferent geographical areas; and ii) a large number of actors operate incircumstances and settings that are respectively very different (Figure 1); andwith iii) different levels of effectiveness.

Since political leadership involves relationships of interdependencebetween leaders, the organization(s) and the context, separately defining‘leader’ and ‘context’ can be limiting. This suggests that trait, contingencyand situational approaches may be usefully complemented by a constitutiveapproach. For example, certain ‘truths’ about political leadership and leadereffectiveness are constructed by the media, by opposition parties, by politi-cal colleagues and by constituents and activists (Grint, 2000). The interpret-ation of particular events such as crises may be inseparable fromconsiderations of leadership (Boin & Hart, 2003). Political leaders alsooperate across interorganizational domains (Hartley, 2002). Though this istrue for (senior) managerial leaders, the influence of political leaders is morefundamental since they govern the institutional and regulatory context foractivity (Moore, 1995). Their activity is also multi-nodal in that they mustconsider the impact of change on different constituencies (Heifetz, 1994).

In illustrating the limitations of an exclusive focus on any of the preva-lent approaches to studying leadership, we have emphasized how under-standing of the individual and understanding of the context are relationallyconfigured, thereby rejecting the dualism implied by casting the individualand the context as separate (Fairhurst, 2001). Different epistemologicalapproaches address the interrelation between individual actions and socialstructures in diverse ways. Rather than dwell on each at length, introducingand outlining some of these briefly is sufficient to position our argument andepistemology. Systems theory (Senge, 1990; Shaw, 2002) emphasizes inter-connectedness, causal complexity and the relation of parts to the whole(Ackoff, 1994). By challenging the notion that social systems can be under-stood in terms of component parts, it frustrates simple causal explanations


of behaviour or outcomes, and underlines the difficulty of predicting theeffects of interventions in social practices. Structuration theory wasdeveloped to address the problem of how to explain the relationship betweenindividual actions and social structures (Giddens, 1984). One mechanism fordoing this is to describe how shared schema (‘scripts’) mediate relationsbetween individual actions and social structures, and influence both recur-sively. Actor network theory (Latour, 1996), and other approaches toanalysing language, such as (D)discourse theory (Honan et al., 2000) orcritical discourse analysis (Potter & Wetherell, 1987), interpret social prac-tices by analysing the way in which texts are presented, and deployed asresources. The privileging of some modes of discourse, and identification ofantinomies can indicate hegemonies or ideologies that pattern relations andactivities, as well as influence the uses and distribution of power in socialgroups (Foucault, 2002; Putnam & Fairhurst, 2001). A host of otherapproaches fall under the broad banner of network analysis, including collec-tive action theory, public goods theory, transaction cost economics theoryand social capital theory (see Monge & Contractor, 2001 for a review).Though varied, what these share is a concern with relationships and inter-actions among a given set of actors, and a focus on social processes as wellas attributes (Katz & Kahn, 1978).

In our model, we use Elias’s term ‘figuration’, which refers to a networkof interdependent actors, and the actors themselves (Dopson, 1997; Elias,1978). Figuration provides a useful, shorthand way of talking about rela-tionally complex domains such as those that characterize the different arenasof politics. More broadly, figurational sociology offers a useful way of inter-preting such relational complexity (Newton, 2002). In comparison with theperspectives sketched above, figurational sociology has some advantages thatsuggest it can be a useful counterpoint to the individual–system dualism inleadership research.

First, figurational sociology is primarily focused on people and theirinterrelationships. As Figure 1 suggests, there are many ways in whichleaders’ interactions may be influenced by the context but an importantfeature of political leadership is the diversity of people with whom theyinteract in a formal capacity, for example: members from the same party atvarious levels; other political leaders; representatives of unions, lobby groupsand communities of interest; business leaders; and individual voters. In eachcase, interactions are part of a wider constellation of interrelationships, andrecursively influence the context for other interactions. This is a more subtlerendering of interaction than economic approaches (e.g. public goods theory,transaction cost economics), which are expressed in terms of relationships ofexchange (Poppo & Zenger, 2002).


Second, figurational sociology, or ‘process sociology’, is concerned withchange, or ‘flux’ (Gouldsblom, 1977 in Dopson, 1997). ‘Systems thinking’and some forms of network analysis are often employed in a diagnosticfashion, used in response to implementing a particular change, to shapefuture strategy, to model a complex process, or as a problem-finding tool(Weisbord & Jaroff, 1995). Similarly, whereas some types of linguisticanalysis, such as critical discourse analysis, are acutely sensitive to the par-ticularities of a given context, other more established traditions withinsociology, such as sociolinguistics, may treat social structures as given, ratherthan as being in the act of created (Putnam & Fairhurst, 2001).

A third, related advantage is that figurational sociology builds thenotion of power into analyses of social networks (Newton, 2002). In anynetwork where actors are not wholly controlled, interdependencies anddifferential power relations produce a complex interweaving of interests andagendas. One implication of this is that search for single causes is futile,which suggests common ground with systems approaches (Senge, 1990).However, a process approach places more emphasis on how such relation-ships are in a state of flux, or becoming. The stress on differential powerrelations, or ‘power chances’, is relevant to understanding political leader-ship, because in democracies, political leaders’ authority is constantly underscrutiny and they are accountable to different constituencies. Political leadershave to build groups and coalitions to address complex problems (Heifetz,1994). Also, their authority is dependent on consent, and open to challengefrom many quarters.

A model of political leadership

In light of this discussion, we propose the following (see Figure 2) as a modelof political leadership. In doing so, we respond to the call for more theoriz-ing in public sector research (Ferlie et al., 2003; Martin, 1997).

By incorporating the contextual filter developed in Figure 1 thiscaptures some sense of the complexities of political leadership. Additionally,it illustrates that leadership is not the province of an individual, but some-thing that is relationally enacted, as represented by the bidirectional arrows.Equally, it incorporates each of the four roles that political leaders variouslycarry out. The first section (roles, responsibilities and KSAs values) followsfrom the institutional and legal framework. Though we have used terms fromlocal political leadership in the UK, we suggest these may be common topolitical leaders in other contexts. The second section (contextual filter)signals contextual contingencies relevant to political leaders, as per Figure 1.


The third section (leadership) shows how leadership may be differentlyconstructed within the different figurations (F1, F2 . . . Fn) in which leadersact. The fourth section (performance) is an outcome measure that showshow leaders’ actions, and the consequences of those actions, subsequently


1

2

3

PERFORMANCE

ROLES Executive Scrutiny Regulatory Advocacy

RESPONSIBILITIES

KSAs VALUES

CONTEXTUAL FILTER

LEADERSHIP

F1 F2

Fn

Figure 2 A model of political leadershipNotes (by section as numbered):1. This is taken as axiomatic and following from the policy literature.2. Contingent, as described in more detail in Figure 1.3. Leadership is socially constituted and relationally constructed in various figurations.

influence future figurations and the contextual filter. Performance could beoperationalized in various ways. We have insufficient space to develop thisin detail, and measuring public sector outcomes is complex and contested(Morrell, 2006) but empirical measures could include electoral turnout,measures of employment, inflation or mortality.

Research agenda

As well as serving a heuristic purpose, Figure 2 can be used to frame aresearch agenda that illustrates how contemporary accounts of leadershipcan be applied in this setting. This capitalizes on the attraction of both visualand propositional forms of theory (Worren et al., 2002). To illustrate this wecan usefully refer back to Grint’s framework of leadership theories. Thoughwe have pointed out the limitations of an exclusive focus on any one of these,the model above can be used to develop a research agenda that usefullycomplements each approach in the following ways.

First, in terms of the different roles leaders undertake, there are differ-ences in the types of generic behaviours that these roles require (Stewart,2003). For example, scrutinizing proposed policy requires a facility forasking the appropriate questions, and an ability to imagine the effects ofpolicy post implementation. Acting as a representative (a function of allpolitical leaders) requires empathy, a willingness and ability to listen, and aservice ethic. Acknowledging the difference between these two roles couldengender a more refined account of what ‘transformational’ behaviour couldmean for example, since similar behaviours across these roles would notnecessarily be equally effective. More specifically, trait or behaviouralapproaches can be encompassed in certain sections of the model. Forexample, research could explore the bidirectional relationship betweenleadership and the contextual filter. One might examine how different leaderbehaviours have an impact on the perceived effectiveness of reform within apublic organization (council, party, national government). Outside the influ-ence on the public body, one could evaluate correlations over time betweenleaders’ style and various outcome measures, such as employment or crime,or perceptions of different stakeholder groups (business, community leaders,minority groups). Ultimately, there may be a relationship between thesebehaviours and (re)election. This goes beyond a straightforward translationof trait-based accounts to this context since it foregrounds the complex,interrelated nature of leadership. This more carefully qualifies the reach oftrait accounts.

Second, in terms of the geographical location, there is reason to believe


that the challenges facing leaders, and their consequent requisite skills differin different localities. This is not simply a case of comparing effectiveness ofleadership in (say) a large, metropolitan area versus (say) a small ruralsetting. Location may be a proxy for a range of differences relating to demog-raphy, technology, social cohesion, etc. So, a contingency approach could beenhanced given the greater scope for modelling diversity encapsulated in the‘contextual filter’. More specifically, contingency approaches could be usedto develop hypotheses about the interaction between political leadershiproles and the context. Similarly, the effectiveness of leadership in an execu-tive function may be a function of the balance of power in a council, or ofthe type of executive arrangement (Leach & Wilson, 2002). This could beexamined further in the diversity of elected mayoral arrangements.

Third, one implication of cybernetics is that models should haverequisite variety in order to study or represent a context or scenario (Ashby,1956). In other words they should have sufficient complexity in their designto enable them to express the range of settings to which they are to be applied(Heylighen, 1992). Figure 1 shows a way in which the diversity of localgovernment can be represented. The challenge of proposing a model of politi-cal leadership is that it needs to have sufficient variety built into it. This is notpossible with existing approaches, but acknowledging the diversity inherentin the domains for political leadership could extend the reach of situationalaccounts. For example, local government could be measured according to thedifferent dimensions and a rating could be given to them across each dimen-sion (demography, community, technology, etc.). These could then beexamined to see how and why political leadership is situationally specific.Additionally, situational approaches would suggest that certain kinds ofleadership behaviour are likely to be more suitable to particular roles (e.g.scrutiny), than others (e.g. representative, or executive leadership).

Fourth, common to trait and contingency accounts in particular hasbeen an emphasis on the individual leader as homo clausus (‘enclosed man’,Elias, 1994), yet as Figure 2 illustrates, elected politicians work with others,or build groups or coalitions. Equally, their authority is dependent onconsent. The model also illustrates how the particular context for a localauthority will: a) influence the interpretation of generic role profiles, and b)influence the type of leadership behaviours deemed effective. The importantthing to note about these two relationships is that they describe a figuration,an interdependent network of actors, and those actors themselves. Eachphenomenon (profile of desired skills, contextual filter, leadership) is rela-tionally constructed, as represented by the thick bidirectional arrows. Sinceleadership itself is a function of interrelations between various actors, this isrepresented by the thin bidirectional arrows. This offers scope to develop


constitutive accounts. One could study councils where there is perceived tobe a ‘strong’ leader, and examine how far contextual factors engender thisperception; or alternatively explore to what extent perceptions of ‘weak’leadership may be attributed to context. Establishing a bidirectional linkbetween the contextual filter and the four generic roles or profiles, andbetween the contextual filter and leadership would provide a validation ofthe constitutive approach, but could also enhance the reach of trait, contin-gency and situational approaches.

Conclusion

The detailed outline of the context for political leaders in Figure 1 offersscope to develop the extant literature on leadership. Trait accounts may becomplemented or enhanced by this outline, since it stresses the importanceof context. Equally, contingency approaches may benefit from alternativeperspectives. Figure 1 illustrates how across various dimensions, a vast rangeof different circumstances within which political leaders act can be modelled.Acknowledging this can enhance the reach of situational models, since thesetypically posit a limited range of scenarios and strategies (Hersey & Blan-chard, 1988; Vroom & Yetton, 1973). Figure 2 details relational complex-ity and underlines the notion that leadership is in flux. Whilst constitutiveapproaches are suited to exploring both these aspects of figurational sociol-ogy, the definition of political leaders, and discussion of particular rolessignal the importance of structural factors in understanding political leader-ship. By stipulating specific dimensions, operationalizing these, and suggest-ing particular behaviours, our approach may complement constitutiveaccounts, offering a clearer empirical agenda for evaluating the ways inwhich leadership is relationally constructed.

This article has offered a glimpse into the world of political leadership.In doing so we provide a model that illustrates the complexity and inherentdiversity of the context for political leaders, addressing calls for furthertheorizing (Leach & Wilson, 2002). The models presented serve a heuristicpurpose, as a summary of the ways in which contexts for political leadershipdiffer (Figure 1), and as a representation of the complex way in whichpolitical leaders’ authority is configured (Figure 2). These also serve to orientfuture research, enhancing the various types of approach to studying leader-ship (Grint, 2000). Trait and behavioural approaches can be used to explorethe relationship between the local context and the type of leadership that iseffective. Contingency approaches can be used to test propositions relatingto the effectiveness of particular behaviours in terms of the role of politicalleaders and their particular context. Situational approaches can be used to


test whether certain behaviours are more or less suitable in terms of politi-cal leaders’ generic roles, and examine leader behaviours across a range ofcontexts. Constitutive approaches can be developed using this model toextend the scope for empirical research through infrequently used paradigms,such as interpretivism and ethnography. Aspects of this approach could beextended to other contexts since, as we have argued, there are some genericsimilarities across different forms of leadership. Studying political organiz-ations can also shed light on generic issues in organizational behaviour(Hartley, 2000). Here, however, our purpose has been to restate the import-ance and coherence of the political realm, allowing scope to explore thiscomplex and substantive area in new ways.

References

Ackoff, R.L. The democratic corporation: A radical prescription for recreating corporateAmerica and rediscovering success. New York: Oxford University Press, 1994.

Alimo-Metcalfe, B. & Alban-Metcalfe, J. The development of a new transformationalleadership questionnaire. Journal of Occupational and Organizational Psychology,2001, 74, 1–27.

Ashby, W.R. An introduction to cybernetics. London: Chapman & Hall, 1956.Azzam, T. & Riggio, R.E. Community based civic leadership programs: A descriptive

investigation. Journal of Leadership & Organizational Studies, 2003, 10, 55–67.Barker, R.A. The nature of leadership. Human Relations, 2001, 54(4), 469–94.Behn, R. Why measure performance? Different purposes require different measures. Public

Administration Review, 2003, 63(5), 586–606.Benington, J. The modernization and improvement of government and public services.

Public Money and Management, 2000, 20(2), 3–8.Boin, A. & Hart, P.T. Public leadership in times of crisis: Mission impossible? Public

Administration Review, 2003, 63(5), 544–53.Boyne, G. What is public service improvement? Public Administration, 2003, 81, 211–27.Boyne, G. & Enticott, G. Are the poor different? The internal characteristics of local auth-

orities in the five comprehensive performance assessment groups. Public Money andManagement, 2004, January, 11–18.

Brunsson, N. The irrational organization: Irrationality as a basis for organizational actionand change. New York: Wiley, 1985.

Burns, J.M. Leadership. New York: Harper and Row, 1978.Chambers dictionary, ed. C. Schwarz. Edinburgh: Chambers Harrap, 1993.Child, J. Organization. Cambridge: Harper and Row, 1984.Collinson, D.L. Dialectics of leadership. Human Relations, 2005, 58(11), 1419–42.Cook, B.J. Politics, political leadership, and public management. Public Administration

Review, 1998, 58, 225–30.Dansereau, F., Graen, G. & Haga, W. A vertical dyad linkage approach to leadership within

formal organizations. Organizational Behavior and Human Performance, 1975, 13,46–78.

Denis, J.L., Lamothe, L. & Langley, A. The dynamics of collective leadership and strategicchange in pluralistic organizations. Academy of Management Journal, 2001, 44(4),809–37.

Dopson, S. Managing ambiguity and change: The case of the NHS. London: Macmillan,1997.


Elias, N. What is sociology? London: Hutchinson, 1978.Elias, N. The civilizing process. Oxford: Blackwell, 1994. (Originally published 1939.)Employers’ Organization and I&DeA. National census of local authority councillors in

England and Wales, 2001.Ensley, M.D., Pearson, A. & Pearce, C.L. Top management team process, shared leader-

ship and new venture performance: A theoretical model and research agenda. HumanResource Management Review, 2003, 13, 329–46.

Fairhurst, G.T. Dualisms in leadership research. In F. Jablin & L. Putnam (Eds), The newhandbook of organizational communication: Advances in theory, research and methods.London: Sage, 2001, pp. 379–439.

Ferlie, E., Hartley, J. & Martin, S. Changing public service organizations: Current perspec-tives and future prospects. British Journal of Management, 2003, 14, S1–S14.

Ferraro, F., Pfeffer, J. & Sutton, R.I. Economics language and assumptions: How theoriescan become self-fulfilling. Academy of Management Review, 2005, 30(1), 8–24.

Fiedler, F.E. A theory of leadership effectiveness. New York: McGraw-Hill, 1967.Foucault, M. The archaeology of knowledge. London: Routledge Classics, 2002.Geddes, M. & Benington, J. Social exclusion and local partnership in the European Union.

London: Routledge, 2001.Ghoshal, S. Bad management theories are destroying good management practices. Academy

of Management Learning and Education, 2005, 4, 75–91.Giddens, A. The constitution of society. Cambridge: Polity Press, 1984.Grint, K. The arts of leadership. Oxford: Oxford University Press, 2000.Grint, K. Problems, problems, problems: The social construction of ‘leadership’. Human

Relations, 2005, 58(11), 1467–94.Hartley, J. Leading and managing the uncertainty of strategic change. In P. Flood et al.

(Eds), Managing strategic implementation. Oxford: Blackwell, 2000, pp. 109–22.Hartley, J. Leading communities: Capabilities and cultures. Leadership and Organizational

Development Journal, Special Issue on Public Sector Leadership, 2002, 23, 419–29.Hartley, J. & Benington, J. Community governance in the information society. London:

Foundation for Information Technology in Local Government, 1998.Heifetz, R.A. Leadership without easy answers. Cambridge, MA: Harvard University Press,

1994.Hersey, P. & Blanchard, K.H. Management of organizational behavior. Englewood Cliffs,

NJ: Prentice-Hall, 1988.Heylighen, F. Principles of systems and cybernetics: An evolutionary perspective. In R.

Trappl (Ed.), Cybernetics and systems. Singapore: World Science, 1992, pp. 3–10.Hogget, P. Conflict, ambivalence and the contested purpose of public organizations. Human

Relations, 2006, 59(2), 175–94.Honan E., Knobel, M., Baker, C. & Davies, B. Producing possible Hannahs: Theory and

the subject of research. Qualitative Inquiry, 2000, 6, 9–32.Javidan, M. & Waldman, D.A. Exploring charismatic leadership in the public sector:

Measurement and consequences. Public Administration Review, 2003, 63(2), 229–42.Katz, D. & Kahn, R. The social psychology of organizations. New York: Wiley, 1978.Kelly, D., Kelly, G. & Gamble, A. (Eds) Stakeholder capitalism. London: Macmillan, 2000.Kets de Vries, M.F.R. The spirit of despotism: Understanding the tyrant within. Human

Relations, 2006, 59(2), 195–220.Latour, B. Aramis, or the love of technology. Cambridge, MA: Harvard University Press,

1996.Leach, S., Hartley, J., Lowndes, V., Wilson, D. & Downe, J. Local political leadership in

the UK. York: Joseph Rowntree Foundation, 2005.Leach, S. & Wilson, D. Rethinking local political leadership. Public Administration, 2002,

80(4), 665–89.Liu, W., Lepak, D.P., Takeuchi, R. & Sims, H.P., Jr. Matching leadership styles with


employment modes: Strategic human resource management perspective. HumanResource Management Review, 2003, 13, 127–52.

Lupton, R. Poverty street: The dynamics of neighbourhood decline and renewal. Bristol:The Policy Press, 2003.

Martin, S. Leadership, learning and local democracy: Political dimensions of the strategicmanagement of change. International Journal of Public Sector Management, 1997,10(7), 534–46.

McKinley, W., Mone, W.A. & Moon, G. Determinants and development of schools inorganization theory. Academy of Management Review, 1999, 24(4), 634–48.

Metochi, M. The influence of leadership and member attitudes in understanding the natureof union participation. British Journal of Industrial Relations, 2002, 40(1), 87–111.

Monge, P.R. & Contractor, N.S. Emergence of communication networks. In F.M. Jablin &L.L. Putnam (Eds), The new handbook of organizational communication. ThousandOaks, CA: Sage, 2001, pp. 440–502.

Moore, M. Creating public value: Strategic management in government. Boston, MA:Harvard Business School Press, 1995.

Morrell, K. Governance ethics and the NHS. Public Money and Management, 2006, 26(1),55–62.

Moss, J. Race effects on the employee assessing political leadership: A review of Christieand Geis’ (1970) Mach IV Measure of Machiavellianism. Journal of Leadership &Organizational Studies, 2005, 11(2), 26–33.

Newton, T.J. Creating the new ecological order? Elias and actor-network theory. Academyof Management Review, 2002, 27(4), 523–40.

Papadakis, V.M. & Barwise, P. How much do CEOs and top managers matter in strategicdecision making. British Journal of Management, 2002, 13, 83–95.

Pfeffer, J. Barriers to the advance of organizational science: Paradigm development as adependent variable. Academy of Management Review, 1993, 18(4), 599–620.

Pollitt, C. The essential public manager. London: McGraw Hill, 2003.Poppo, L. & Zenger, T. Do formal contracts and relational governance function as substi-

tutes or complements. Strategic Management Journal, 2002, 23(8), 707–25.Potter, J. & Wetherell, M. Discourse and social psychology. London: Sage, 1987.Pugh, D.S. The Aston program perspective. In A. Van de Ven & W. Joyce (Eds), Perspec-

tives on organization design and behaviour. New York: Wiley, 1981, pp. 135–65.Putnam, L.L. & Fairhurst, G.T. Discourse analysis in organizations. In F. Jablin & L.

Putnam (Eds), The new handbook of organizational communication: Advances intheory, research and methods. London: Sage, 2001, pp. 78–136.

Rallings, C., Temple, M. & Thrasher, M. Participation in local elections. In L. Pratchett &D. Wilson (Eds), Local democracy and local government. London: Macmillan, 1996,pp. 62–83.

Ranson, S. & Stewart, J. Management for the public domain: Enabling the learning society.London: Macmillan, 1994.

Robbins, B. (Ed.) The phantom public sphere. Minneapolis: University of Minnesota Press,1993.

Sanders, J.E., Hopkins, W.E. & Geroy, G.D. From transactional to transcendental: Towardan integrated theory of leadership. Journal of Leadership & Organizational Studies,2003, 9(4), 21–31.

Senge, P. The fifth discipline. New York: Doubleday, 1990.Shaw, P. Changing conversations in organisations: A complexity approach to change.

London: Routledge, 2002.Skelcher, C. & Mathur, N. Governance arrangements and public service performance.

Birmingham: University of Birmingham, Inlogov, 2004.Sparrow, R.T. & Liden, R.C. Process and structure in leader member exchange. Academy

of Management Review 1997, 22(2), 522–52.


Stewart, J. Modernising British local government. London: Palgrave, 2003.Thompson, J.D. Organizations in action. New York: McGraw-Hill, 1967.Vera, D. & Crossan, M. Strategic leadership and organizational learning. Academy of

Management Review, 2004, 29(2), 222–40.Vroom, V.H. & Yetton, P.W. Leadership and decision making. Pittsburgh, PA: University

of Pittsburgh Press, 1973.Weisbord, M.R. & Jaroff, S. Future search: An action guide to finding common ground in

organizations and communities. San Francisco, CA: Berret-Koehler, 1995.Worren, N., Moore, K. & Elliott, R. When theories become tools: Toward a framework

for pragmatic validity. Human Relations, 2002, 55(10), 1227–50.Yukl, G.A. Leadership in organizations, 2nd edn. Englewood Cliffs, NJ: Prentice Hall,

1989.Zald, M.N. More fragmentation? Unfinished business in linking the social sciences and the

humanities. Administrative Science Quarterly, 1997, 41, 251–61.


Kevin Morrell is Senior Research Fellow at the Institute of Governanceand Public Management,Warwick Business School,University of Warwick.Dr Morrell’s interests centre on behaviour within organizations. To datehe has carried out research into employee turnover, ‘leadership’, ethicsand value in organizations. The empirical context for his research hasbeen the public sector, viz. the National Health Service and UK localgovernment. As well as drawing out and substantiating the widerimplications of this research for other sectors, he has written cognatetheoretical papers that are not sector specific. Before joining WarwickBusiness School, he was an ESRC postdoctoral fellow at Loughborough.[E-mail: [email protected]]

Jean Hartley is Professor of Organizational Analysis, Institute ofGovernance and Public Management, Warwick Business School,University of Warwick. She is an organizational psychologist, whose workis concerned with political and managerial leadership, and organizationalinnovation and improvement in public services, particularly central andlocal government, health and the criminal justice system. She hasdeveloped a tool for assessment and development of formal politicalleadership, called the Warwick Political Leadership Questionnaire(WPLQ). Professor Hartley has experience of evaluation research forcentral and local government. She is currently the Research Director forthe research that is monitoring and evaluating the Beacon Schemeconcerned with interorganizational learning, innovation and improve-ment. She has recently been Lead Fellow of the ESRC/EPSRC AdvancedInstitute of Management Public Services Group.[E-mail: [email protected]]

a model of political leadership - control...

Documents