Causes versus consequences, processes versus outputs in social

Heiner Meulemann
Forschungsinstitut für Soziologie, Universität
zu Köln
Greinstraße 2, D50939 Köln
Tel. 0221 - 470 5658, Fax 0221 - 470 5169
e-mail: [email protected]
Perspectives
on
Social
Capital
> traditional-corporatist regime
Definition, questions and some results from the
European Social Survey
Lecture at the X. Conference of the SU-Higher School of Economics,
Moscow, April 6-10, 2009
I like to express my gratitude to the
“Verein der Freunde und Förderer der Universität zu Köln”
(Association of friends and sponsors of the University of Cologne)
for the generous support of this research visit.
1
Three topics
1 How is social capital (SC) best defined?
According to this definition:
2 Which questions on SC should be
reserarched first? And which have? First
priority: Transfer hypothesis
3 Example of transfer hypothesis
- European social survey
- Economic sphere of labour relations
2
1 Defining social capital:
Social relations as resources
of actors in contexts
3
1.1 Criteria
• Putnam: “refers to features of social
organization, such as trust, norms and
networks”
• Combines reference to social process –
“organization” – with enumeration – “such
as”.
• I will analyze what is “social” of “features
of organization” in order to understand the
enumeration
4
„Features of social organization“
= collective good of organizations. Organizations
can be distinguished by their SC – just as by
constitution, function, size. Grammatical singular
“organization” accidental.
Yet: singular meaningful: “organization” = process
sustained by persons. Members of a group
interact, “organize themselves”, so that collective
products result
Question: what “collective goods” or products?
5
Three collective goods,
resulting from interaction of persons
(1) productivity of a network: from flow of exchanges
between persons with positions and intentions
(2) climate of trust: sufficient number of people reciprocate
benevolent actions; if number goes down, trust risky,
vicious circle: climate of distrust
(3) validity of norm: sufficient number follow norm and
sanction violations; if number goes down, vicious circle
of deviance and tolerance, norm breaks down
In each case: “collective good” from interaction of
members. “Features of organization” established bottom
up
Therefore, genus proximum of definition not organizations,
but persons.
New definition 1: SC = any property of a group member,
which contributes to group outputs.
6
Problem of new definition 1:
too broad, „social“ lost
Contribution of persons also from human or
cultural capital
Therefore restricton: only outputs from
membership in group.
In pursuing common interest, members form social
relations, interactions more densely knit
amongst members than with non-members.
Somewhat narrower definition 2: SC =
sum of social relations a person holds in
groups
7
Problem of new definition 2: still too broad,
includes intimate relationships
Intimate = sexual and generational relations, rest
on biology; everybody can, and most will, enter
into them. Practiced in “private living
arrangements”.
“Private” = “particularistic” (Parsons): person
essential for the relation
many mothers, but only my mother is “my” mother
SC consists of “universalistic” relations in “public”
realms, persists with interchangeable persons
New and final definition 3: SC = sum of social
relations a person holds in groups beyond
intimate living arrangements
8
Uses of SC in social contexts
SC does not “capitalize” by itself. In order to
not decay, it must be utilized. What can
actors gain from SC?
Due to its relational nature, SC more useful
when more relations in group.
Therefore, distinction of contexts of use:
9
1.2 Relational capital and system capital:
Concept and measurement
“Relational SC” of persons - “system SC” of group
System SC: conceived of independently of persons as sum
of relations.
Person may aim to manipulate and to improve relational SC
System SC of group exists independently of members
Emerging quality in two ways
(1)network of relations knitted between members in
pursuance of group goal = social structure of the group.
(2)Some relations bundled in civic associations. System SC
= sum of civic associations acting within group
Measured in surveys by aggregation
Two problems:
10
Measurement problem 1:
interdependency, solitary decisions
In social reality interdependency: some are eager to and
some detest emulating other people.
Measurement should follow up interdependency until SC is
established
Yet surveys
- neglect interdependency and time
- Instead: means within groups at single time point
Justification: Practical short-cut?
Yes, but also substantive reasons: some decisions made
without looking at others (join a tennis club).
If this holds: measure of system SC as group mean of
relational SC also theoretically justified
11
Measurement problem 2:
circularity, random sampling
Danger of circularity: system SC = relational SC. Yet:
Sum of relations of all persons does not necessarily
amount to system SC of group.
Example 1: Two persons related = only one relation,
counting two relations not correct. However, random
samplings: improbable that two persons with relation are
drawn. Total as system SC feasible
Example 2: Two persons join same association = one, not
two association. Again, random sampling. Furthermore:
multiple memberships reflect size of association. Again,
total as system SC feasible
12
Summary so far
SC consists of relations of persons, basically
relational
Relations add up within a group to system SC:
(a) network, social structure
(b) civic associations
Although system SC conceived of as independent
of relational SC, measurement of system SC
through mean of relational SC can be justified.
Mean of relational SC = indicator of system SC.
Question: Which properties of system SC fruitful
for group member in pursuance of goals?
13
1.3. Three Properties
of System SC
From (1) density of social relations
to
(2) social trust and
(3) validity of norms
(1) fundamental, (2) and (3) derived
This to be shown in following
14
(1) Density of social relations
Network of high mean personal relationships eases moves,
each partner has more relations. That is: value of
relational SC increases with system SC
Quantity of relations increases the number of ways to attain
goals
Quality of relations affects probability of attainment on
these ways. In particular:
“Niceness” eases goal attainment. Starts interactions with a
cooperative move, and end up better than people
starting with a non-cooperative move (Axelrod)
“Nice” relations result from two “nice” tendencies of
partners
- to trust each other, > 2nd property
- to endorse norms of cooperation, > 3rd property
15
(2) Climate of trust
Trust = overrides suspicion that partner will not give back.
More trusting, longer chain of reciprocation, stronger
climate of trust
Trust learned in “particularistic” relations, reinforced
reciprocally in “universalistic” interactions
I trust in others who have repeatedly not disappointed me,
and others trust in me if I have repeatedly not
disappointed them.
My trust in others indicates the trust others have in me.
Trust not personal attitude alone, but indicator of trustful
relations in group.
If trust = indicator of “niceness” of relations,
climate of trust = system SC useful for persons
16
(3) Validity of norms of cooperation
Norms of cooperation (proscription “not to”), justified by the
norm of reciprocity alone.
Norms of institutions (e.g. marital fidelity), additionally
justified by values the person beliefs in (“family” or “life”)
Endorsement of norms of cooperation, more or less
strongly reciprocated:
- Behavior: If enough follow norms of cooperation and
enough sanction violations, norm becomes valid.
- Attitude: If enough endorse norm, it becomes valid
Consequently, the more norm held among interaction
partners, the more one can uphold norm oneself.
Endorsement of norms = indicator of “nice” relations
Validity of norms of cooperation = system SC
17
In sum: Triad of system SC, but
only „relations“ relational SC
Distinction between density and “niceness” of social
relations justifies to classify system SC into Putnam’s
triad: “networks, trust, and norms” (order changed!).
Yet:
- density of relations only justified directly as a system SC
- further arguments required to classify climate of trust and
validity of norms of cooperation as system SC: indicators
for “niceness” of social relations, not measured directly
with reference to relations, but indirectly as means of
attitudes.
Test: switch back from system SC to relational SC:
- Just as density of relations = system SC, so relations of
person = “relational” capital.
- However, while climate of trust and validity of norms =
system SC, trust or norm endorsement not = SC of
18
person.
1.4 System SC as context:
Social order and opportunity structure
Which kinds of groups bearer of system SC? Any
aggregation level or “context”.
Trivial question? No. To explain system SC, reference to
analytical properties of groups. Question changes:
Which kinds of analytical properties define their system
SC? Two:
(1) Name and a border, constitution and laws, folklore and
customs. Become “social facts” = social order.
(2) Resources of action: money, educational degrees,
power. Circulate among citizens and make up different
“life chances” = opportunity structure.
19
Context 1: Social Order
Consists of norms directly guiding actions. Valid because
- most people endorse them
- in large parts, written down in legal form: constitution.
Example: equality defined in constitutions, achievement (equality’s twin
value) only in peoples’ minds
Typical variables: federal or unitary constitution, % Protestants
(tradition of self-determination)
Guides actions in same way as personal endorsement of
norm – only difference: binds every citizen
Must be symbolically identified: kings or presidents, laws
and customs, flags and hymns, border stones and
national football teams.
Higher aggregation level = more important. Nation state
has a social order, city precinct not
20
Context 2: Opportunity Strucuture
Sets de facto range of options for every citizen, beyond
personal resources, indirectly affects actions.
Results from actions of all citizens and all organizations of
country.
Examples: Social inequality, reduces trust. Democracy since long,
facilitates associations.
Options and restrictions in same way as opportunity profile
of person (combination of resources) - only difference:
for every citizen alike
Need not be symbolically identified
Lower aggregation level = more important level. Public
swimming pool in neighborhood, not in city.
21
Cross-Classification
with societal domains
Social order
Opportunity
Structure
Economy
Economic
GDPpc
Freedom (EFR)
Polity
Good
Governance
Years of
democracy
22
2 2 Classifying and
evaluating research
questions on SC
23
“Capital” two qualities
(1) every capital = means for ends to be
attained in purposive action
(2) every capital “capitalizes” = pays off in
same kind
These two qualities = dimensions to classify
research questions
24
Quality 1: means to ends in
purposive action
Money = exchange against goods and services.
Prestige = used to attain goods and services from others.
SC = channels to goods and services.
Each: means to “success”. Yet difference:
Money buys everything of its worth. “Success” no problem.
Prestige, SC: be worked upon to become a means. “Success” problem.
Thus: If SC contributes to success, consequences should
before causes. If not, reduced importance of causes.
SC research agenda: 1 consequences - 2 causes
Money “success” for everyone who holds it, first: how got it;
second: what done with it.
Money research agenda: 1 causes - 2 consequences 25
Quality 2: capitalization
Money = interest.
Prestige of educational degrees = prestige of occupations
attained
SC = social relations, pay off in social relations. That is:
- Relational SC the more useful, the more embedded in
network of relations, the more system SC.
A’s relation to B
= limited value if B knows nobody,
= highly valuable if B at core of network.
- Due to relational nature of SC, capitalization depends on
context
Research agenda: priority of effects of system SC:
1 on a means end chain of some action (“slopes”)
2 on ends themselves (“intercepts”)
26
Figure 1: Causes versus consequences,
processes versus outputs in social capital
research
CONSEQUENCES
CAUSES
Context: System Capital, others
Context: Others
3 Social Capital
as result
3 Formation
b
Resources/
Attitudes
1 Capitalization
a
Social Capital:
end
b
Social Capital:
means
2: Output of social capital,
a
Action goals
a = effect on (interceot, independent variables centered=) mean, mean hypothesis
b = effect on (slope=) relation, effect hypothesis
(a,b effects on the corresponding parameters of micro level regression)
27
Agenda followed by research up to now?
Yes, but only implicitly
Research not in SC per se or causes, but in consequences
for social integration, democratic stability of nation state.
At heart of SC research: transfer hypothesis. “Good
government is a by-product of singing groups and soccer
clubs” (Putnam). Abstractly: citizens’ involvement grants
social integration.
Transfer hypothesis: on consequences, implicitly priority of
consequences over causes. But apart from that, not
clear.
Meaning specified: using right half of figure 1.
.
28
Transfer hypothesis, specified
Country level correlation
Corresponding person level effect:
= more citizens in associations, more articulation of interests in democratic
decision making
= Transfer from civic life to organized social life.
= Figure 1: SC as a means > action success
Two Problems:
(1) Articulation of interests not yet social integration. Further causal link
from successes of persons to integration of groups, ultimate impact
of social capital research. Beyond figure 1 to the right. Mostly, taken
for granted on theoretical grounds and not researched empirically.
(2) Reference also to embeddedness in macro conditions. Thus, “the
singing groups and soccer clubs” = cipher for system SC. But its
effects on micro relation not specified. Therefore: both effects of
figure 1
29
Transfer hypothesis, summarized
Comprises right half of figure 1 as a whole and
expands it to the right. To be tested, its four
elements must be specified
(1) Macro relation. To which added
Two top down elements:
(2) Effect hypothesis, capitalization of system SC
(3) Mean hypothesis, output of system SC.
A new bottom up element:
(4) From action success to social integration
30
3 Example:
Empowerment at the
work place
31
3.1 Question and research
design
Transfer of transfer hypothesis
From politics
The more someone is involved in private
associations, the more..
- able to assert political interests
To labor relations system
- attain empowerment at the workplace
= range of discretion in order to make
decisions about work
32
Controls to examine transfer
hypothesis of labor relations
On the level of persons
- Human capital, union membership,
workplace
On the level of countries
- Institutions and opportunity structures of
labour relations system
33
Research Design: Influences on
empowerment
5 Collective: work place
sector, size of firm
?
4 Collective: strategy
union membership,
?
+
+
+
(+)
1 Civic Involvement
social capital
(+)
2 Human Capital in Firm
People supervised,
Prestige of occupation
(+)
3 Human Capital, person
political efficacy, education,
Exit options
Empowerment:
discretion at work
+
+
-
+
Labour relations system:
Favourable to unions
34
Dependent Variable:
Index of Inventory and a Question
Inventory: “Please say how much the management at your
work allows you:
- to be FLEXIBLE in your working hours,
- to DECIDE how your own daily work is organised,
- to influence your work ENVIRONMENT,
- to influence decisions about the general DIRECTION of
your work,
- to CHANGE your work tasks if you wish to?”
“0 I have no influence” - “10 I have complete control”
Single question: “To what extent can you ORGANIZE your
own work, to a large extent (4), to some extent (3), very
little (2) or not at all (1)?” – reversed for analysis
Sample: ESS 2002, employed population
35
Figure 1. Mean empowerment, one standard
deviation above and below means
10
10
9
9
8
8
7
7
Highest N: 6.69
6
6
5
5
4
Lowest PL: 2.49
4
3
3
2
2
1
1
0
0
A
B
DK
E
FIN GR
IRL
IT
L
N
NL
P
S
UK D-W D-O H
PL
SLO
36
Empowerment
at the Work Place
High: Scandinavian countries, NL
> social democratic regime
Medium: AU, B, EI, I, LUX
> traditional-corporatist regime
Low: E, GR, PT, D-E, D-W
> traditional-corporatist regime
37
3.2 Hypotheses and measurements:
Level of persons
(1) Civic Involvement: +
(2) Human capital: Person: +
Political efficacy, education, exit options
(3) Human capital: Firm specific: +
People supervised, prestige of occupation
(4) Union membership: +
(5) Work Place
Size, sector of firm
(6) Control variables: Age and Gender
38
(1) Civic Involvement
in private organizations
“For each of these voluntary associations, tell me whether any of these
things apply to you now or in the last 12 months
- A member of such an organization
- Participated in an activity arranged by such an organization
- Donated money to such an organization
- Done voluntary (unpaid) work for such an organization.”
Membership + participation = belonging
Donation of money + voluntary = engagement
In 5 private associations:
- sports clubs
- consumer associations
- scientific/educational/teachers’ associations
- social clubs
- cultural associations
39
(3) Exit options
- Index of (1) “How difficult or easy would it
be for you to get a similar or better job with
another employer?” and (2) “... and to start
your own business?”, scale 0 to 10: +
- partner employed: +
40
3.3 Hypotheses and measurements:
Level of countries
Countries characterized by labor rule system
LRS, more or less favorable to workers
= rule set and power structure, which
exonerate workers from personal
endeavor to attain empowerment; if unions
successful = Workers less dependent on
their own initiative
41
Two Dimensions of LRS
Dimensions
- Relations between collectivities or individuals
- Regulation through normative social order or factual
opportunity structure
Four-Fold Table
Collective relations
- Normatively: range covered by bargaining process
- Factually: a high degree of organization and public
support
Individual relations
- normatively: favor employment and restrict dismissal
- Factually: a labor market situation with high employment
and many secure work contracts
Indicators in following table
42
Table 1 Variables of the labor
relation system
43
Mean hypothesis
Union efficiency hypothesis:
The more collective or individual labor
relations of a country favor
normatively, or strengthen factually,
the unions,
the higher mean empowerment of
employees
44
Effect hypothesis
Substitution hypothesis:
The more the labor relation system of a
country favors unions, the less important
individual strategies become for the
worker in order to attain empowerment.
Negative cross-level interaction effect
between LRS favorable to unions and
individual endeavor to attain
empowerment, in particular: human capital
45
3.4 Results
Mean Union membership:
- .359 mean of 19 countries
- Range from .146 Portugal
to .844 Denmark
46
Table 3 Multi-Level-Regression of Empowerment on Person and
Country Variables: Raw Coefficients
Variables (Categories)
Empty
Intercept
Mean Intercept
5.233
Union Membership mean
Civic Involvement
Private-Belonging
Private-Engagement
Firm-Specific Human Capital
People Supervised (5)
P Superv * UM mean (*100)
Prestige (*10)
Individual Strategy: Personal Human Capital
Political Efficacy (5)
Education (7)
Exit Option (11)
Exit O. * UM mean (*100)
Partner Employed
Collective Strategy
Professional-Belonging
Workplace
Firm size (5)
Sector: Service
Control Variables
Male
Age (b for 10 years)
Random
Intercept
Intercept
Outcome
Intercept+
2 Slopes
Intercept+ 1
Slope
5.406
5.472
.019***
5.509
.019***
5.512
.019***
.242
.207
.240
.205
.227
.203
.233
.203
.445
.445
.445
.282
.284
.456
-.298
.283
.143
.139
.213
.143
.137
.213
.124
.123
.144
.132
.216
-.174*
.126
.143
.134
.217
-.198**
.124
-.203
-.210
-.207
-.205
-.191
(.064)
-.191
(.065)
-.192
(.063)
-.189
(.063)
(-.125)
.160
(-.125)
.160
(-.128)
.160
(-.128)
.160
.283
47
Table 3 Multi-Level-Regression of Empowerment on
Person and Country Variables: Variance Components
Variables (Categories)
Intercept
Variance Components
Person-Level
Country-Level: Intercepts
Empty
5.047
.616
Random
Intercept
3.809
.285
Intercept
Outcome
Intercept+ Intercept+ 1
2 Slopes
Slope
3.810
.129
3.774
.111
Slope1: P Supervised (*100)
.689
Slope2: Exit Option (*100)
.331
Intercept*Slope1-Correlation
-.760
Intercept*Slope2-Correlation
-.383
Slope1*Slope2-Correlation
Intra-Class-Correlation
R2 Persons
R2 Countries
Deviance
Df (Deviance)
N of persons
3.784
.114
.332
-.416
-.187
.1094
69411
2
15333
.0696
.245
.538
60383
2
0.033
.245
.791
60380
2
14429
.252
.819
60295
7
.250
.815
60315
4
48
Quantity of
intercept and slope effects
Mean:country 25 percentage points above
grand mean > predicted intercept of 5.472
+ 0.019*25=5.947, half a point on 11 point
scale of empowerment.
Slope: country 25 percentage points above
mean > predicted slope for
- people sv.: .456 + (-.00298*25) = .382
- exit option: .216 + (-.00174*25) = .172.
49
3.5 Conclusion:
Hypotheses confirmed?
Transfer hypothesis: confirmed. Effects stronger than many
effects of the more immediate personal factors
- Belonging to and engagement in private associations
stronger than belonging to trade unions. Articulation of
one’s interest in private realm more easily transformed
into empowerment at the workplace than in public realm.
- More distant = more effective route. Longer distances
needed to acquire general capacities of self-assertion,
while focusing on the very arena of interest narrows
down opportunities to learn general capacities. Longer
distances = more challenges to generalize.
50
I like to express my gratitude to the
“Verein der Freunde und Förderer der
Universität zu Köln”
(Association of the friends and sponsors of
the University of Cologne)
for the generous support of this research
visit.
51