The Fractal PolityWhy Schröder Failed at the Second Term

Haiko LietzDepartment of Sociology

Columbia Universityhl@haikolietz.de

May 20, 2007

NetSci 2007International Workshop and Conference on Network Science

Introduction

A Question of Governance

Two years ago: Red-Green government coalition under chancellor Gerhard Schröder failed during the second term

Agenda 2010: Shift of governance in labor policy domain from state to market transactions

Situation: Narrow gov’t majority in lower house, strong opposition majority in upper house

How can we explain Schröder’s failure of governance? (by just dealing with the lower house)

2

Introduction

3

How can we explain Schröder’s failure of governance?

Success and failure = Policy outcomes

How is policy outcome produced?

Outcome = Action

How does social structure constrain action?

What is social structure?

Emergence and self-similarity are the keys

Neither “Network” nor Hierarchy

Governance school (Kenis & Schneider 1991, Marin & Mayntz 1992, Börzel 1998)

Policy networks as a form of governance between market and hierarchy

Criticism: implies that markets and hierarchies are not networked processes

Interest intermediation school (Marsh & Rhodes 1992)

Policy networks as meso-level analytical concept for institutionalized relations between state and civil society

Criticism: implies that micro-level of individual interaction and macro-level of social structure are not networked processes

Networked Polity (Ansell 2000)

Networked polity is a complex system of policy interest intermediation that embeds networked organizations which are themselves complex systems

Heterarchy (Hedlund 1994, Stark 1999): complex adaptive system of organization

4

Heterarchy as Alternative Form of Governance to Hierarchy

5

Hierarchy Heterarchy “Network” Market

Communication asymmetrical symmetrical symmetrical both

Aggregation ofComponents

centralized decentralized decentralized none

Self-organization

no yes yes none

Organization ofDiversity

yes yes no none

Form ofGovernance

yes yes no no

Identity and Control

6

Structural theory of social action (Harrison White Forthcoming)

A systemic perspective on social organization

Interaction

… – Atoms – Molecules – Cells – Tissues – Organs – Organisms –

biophysical

Components

social

… – Persons – Groups – Communities – Crowds – …

Feedback / Constraint

Higher levels emerge out of, and constrain, component interactions

Identity and Control

Central concepts

Identities are the social components

Identities seek control (“footing”) in stochastic environment, amidst contenting control projects

Control by avoiding social stress (mismatch)

As identities settle and cluster they embed into “social molecules” called “disciplines”

Disciplines steer identity interaction

7

Identity and Control: Social “Arena” Molecule

Deterministic model first proposed by Barabási, Ravasz & Vicsek (2001) 8

Steers selection processes

Reduces social stress by severing ties to “impure” identities

Example: homophily inside political parties

Identity and Control: Social “Interface” Molecule

Deterministic model first proposed by Barabási, Ravasz & Vicsek (2001) 9

Steers flows

Identities commit to ensure quality of flow

Example: formal governance (hierarchy)

Identity and Control: Social “Council” Molecule

Deterministic model first proposed by Barabási, Ravasz & Vicsek (2001) 10

Steers mediation processes

Identities interact on the basis of prestige

Example: informal governance (heterarchy)

Identity and Control

Central concepts (cont’d)

Institution: a social process taken for granted, evolved and stable social structure (skeleton of social structure)

Style: profile of uses of institutions, “sensibility” towards other identities and their styles (heartbeat of social structure)

Identities seek control by enacting styles on different levels

Social structure is social space-time: Process is structure, structure is action

Persons already are complex emergent forms

Social structure has a fractal dimension

11

Agenda: Understanding of something that needs to be achieved

Success: Enactment of style, implementation of agenda

Levels of Embedding

Person

Party

Parliament

Coalition / Opposition

Democracy

Problem: If styles are enacted rigorously, orderly structure will block action

Solution: “Getting action must continue to break u the hardening crust of issues and interests which congeal continually to block action.” (White Forthcoming: 7.3.2.)

Agenda and Successful Governance

12

Complexity / Stress Possibility to enact style

Agenda and Successful Governance

Successful governance: implementation of gov’t agenda, embedding of style in the institution of legislation

Sufficient condition: enactment of gov’t style by coalition deputies

Necessary condition: heterarchical self-organization of coalitional and oppositional diversity to exploit the friction

Hypothesis: Either condition was not fulfilled in unsuccessful policy domains

13

Research Design: Data for 2002-2005 Legislative Period

14

Table 1: Number of affiliations of deputies in committees and non-parliamentary organizations

SPD CDU/CSU Green FDP PDS Total

Deputies N 221 244 43 44 2 554

Affiliations*

Committees 532 (18) 496 (16) 116 (4) 94 (4) 2 (0) 1240 (42)

Publicorganizations

192 (166) 249 (235) 30 (26) 33 (30) –– 504 (457)

Privateorganizations

877 (410) 686 (519) 145 (101) 150 (92) 6 (2) 1864 (1124)

Businesses 152 (148) 303 (301) 29 (29) 55 (54) –– 539 (532)

Unions 128 (2) 5 (0) 9 (0) –– 2 (0) 144 (2)

Total 1881 (744) 1739 (1071) 329 (160) 332 (180) 10 (2) 4291 (2157)

Notes: Numbers in brackets are primary affiliations. Table shows only data for deputies that serve in committees. Public organizations: ministries, statutory corporations, agencies, etc.; private organizations: associations, clubs, foundations, think-tanks, etc.

* p < 0.01 (qui-square test for all affiliations except committees)

2-mode network of all 601 deputies and their affiliations

Color codes: SPD (red), CDU (black), CSU (blue), Greens (green), FDP (yellow), PDS (orange). Grey nodes are organizations. Size of nodes is degree. 15

Results

Scale-free property Small-world properties

Power-law degree distribution for the full parliament.

Scaling exponent is 0.94. R-squared (adjusted) is 0.74.

Clustering Coefficient and Average Distance for

the substructure of the full parliament

16

0

1

2

3

4

5

6

0 1 2 3 4 5

ln (degree k)

ln (

# o

f n

od

es w

ith

deg

ree

k)

0

0.2

0.4

0.6

0.8

1

1.2

3 4 5 6 7

ln (size)

clu

ster

ing

co

effi

cien

t

1

1.5

2

2.5

3

3.5

4

4.5

3 4 5 6 7

ln (size)

aver

age

dis

tan

ce

Research Design

Foreign Policy Domain: successful, mandate of armed forces repeatedly extended in 2002, 2003, and 2005 by large majorities in coalition and opposition

Labor Policy Domain: induced failure of whole government, caused demonstrations and “bad press”

Organizations selected that were relevant to the domain (Laumann & Knoke 1987)

2-mode networks are transformed into directed and undirected 1-mode networks (Breiger 1974):

17

Affiliation of deputy B

Primary Secondary

Affiliationof deputy A

Primary A↔B A→B

Secondary A←B A↔B

Results: Foreign Policy Domain (N = 140)

Prime example of heterarchical organization

Both coalition and opposition well positioned to organize their agendas: high densities from organizations in principal agreement with the gov’t agenda

18

E-I Index (Krackhardt & Stern 1988):

E-I = ( bridges – non-bridges) / realized ties

Between -1 (full polarization) and 1 (no internal ties)

Results: Labor Policy Domain (N = 95)

Coalition deputies mass-affiliated to unions and workers milieu org’s

Coalition deputies’ agenda differed from gov’t agenda

Highly negative E-I score

19

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

3 4 5 6 7

ln (size)

E-I

in

dex

Env. Policy Domain (undir.)

Env. Policy Domain (dir.)

Foreign Policy Domain (undir.)

Foreign Policy Domain (dir.)

Health Policy Domain (undir.)

Health Policy Domain (dir.)

Labor Policy Domain (undir.)

Labor Policy Domain (dir.)

Conclusion

Initial question: How can we explain Schröder’s failure of governance?

Social structure emerges from, and constrains, the enactment of style by identities across different levels

Avoidance of stress results in social “order” that comes to block action

Fresh action necessarily confronts and exploits diversity: heterarchy

Agenda 2010 contradicted the agenda of those deputies that were supposed ti implement it

At same time Red-Green gov’t was successful: other policy domains

Promising results, but more data needed

20

Heterarchy is multi-level concept:

Person (White Forthcoming)

Firm level (Stark 1990, Beunza & Stark 2004)

Industrial production system (Nishiguchi & Beaudet 2000)

Economy (Stark 2001)

Urban governance (Davies 2005, Kearns & Paddison 2000)

EU governance (Jessop 1998, Neyer 2003, Scott & Trubek 2002, Smismans 2006)

Parliament

Our institutions are organized hierarchically:

Parties, parliaments

The sociological concepts we use are self-similar:

Discipline, style

It is a fractal polity

The Fractal Polity

21

The Fractal Polity

Real networks well described by deterministic models (Ravasz & Barabási 2005)

Construction: adding quadrads between modules: weak heterarchical organization

Hierarchy of correlated hubs and community structure emerge (Newman & Park 2003)

22