Rethinking the Decline in Social Capital

• Author: April K. Clark
• Published date: 01 July 2015
• Date: 01 July 2015
• DOI: http://doi.org/10.1177/1532673X14531071
• Subject Matter: Articles

531071APRXXX10.1177/1532673X14531071American Politics ResearchClark
research-article2014
Article
American Politics Research
2015, Vol. 43(4) 569 –601
Rethinking the Decline in
© The Author(s) 2014
Reprints and permissions:
Social Capital
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1532673X14531071
apr.sagepub.com
April K. Clark1
Abstract
This study investigates trends in social capital in the United States since the
1970s. The literature suggests that variations in social capital are associated
with both individual attributes and macro-level economic conditions. Yet,
others argue that after controlling for these features, large-scale changes
in social capital are evident across birth cohorts and over time. While
previous studies have identified a number of individual and societal factors
that influence social capital, I note that the common modeling approach
used is inappropriate for examining the interaction between national and
individual-level data as well as the simultaneous influence of period- and
cohort-based effects. I therefore utilize a multilevel model to reassess the
different theories of the origins and determinants of social capital. The
evidence presented casts doubt on past studies that see a general erosion
in social capital as well as those that view the decline as stemming from
generational replacement.
Keywords
social capital, determinants, period, cohort, trends
Introduction
In the last 20 years, concerns about a decline in social capital and the accom-
panying detrimental effects on civic culture have spawned an immense
amount of literature. The benefits of social capital, defined as the connections
1Purdue University Calumet, Hammond, IN, USA
Corresponding Author:
April K. Clark, Assistant Professor of Political Science, Department of History and Political
Science, Purdue University Calumet, Classroom Office Building #215, 2200 169th Street,
Hammond, IN 46323-2094, USA.
Email: april.clark@purduecal.edu.

---

570
American Politics Research 43(4)
among individuals, and the social networks and norms of reciprocity that
arise from them (Putnam, 2000) are by now well documented. Social capital
is associated with economic growth (e.g., Fukuyama, 1995; Knack & Keefer,
1997), greater levels of income equality (e.g., Kawachi, Kennedy, Lochner,
& Prothrow-Stith, 1997; Wilkinson, 1996), higher education (e.g., Coleman,
1988), lower crime rates (e.g., Putnam, 2000; Wilson, 1987), better commu-
nity welfare and public health (e.g., Kawachi et al., 1997; Wilkinson, 1996),
and more responsive government (e.g., Putnam, 1993). Thus, a decline in
social capital in American society is fundamentally important as research
suggests that social capital brings good things politically, civically, and eco-
nomically. In short, it is understood that high levels of social capital are
imperative for a functioning democracy (Coleman, 1990; Fukuyama, 1995;
Paxton, 1999; Putnam, 1993, 1995a, 2000; Uslaner, 2000-2001).
While the importance of social capital is widely accepted, considerable
debate remains surrounding trends in social capital.1 While Robert Putnam
(1995a, 2000) in his seminal article and subsequent book-length study argues
that there is a general decline in social capital in the United States, others
contend that participation levels and social cohesion isn’t declining and so
there is no cause for concern about the country’s civic life (Costa & Kahn,
2003b; Ladd, 1996, 1999). What is more, disagreements exist as to the spe-
cific source of change in social capital. Social capital levels in the United
States may change over time, reflecting a period effect, or across generations,
reflecting a cohort effect. Previous studies have argued that large-scale
changes in social capital are evident across birth cohorts with the replacement
of more civically-minded generations with newer, less engaged and trusting
generations (Brehm & Rahn, 1997; Clark & Eisenstein, 2013; Paxton, 2005;
Putnam, 1995a, 2000; Robinson & Jackson, 2001; Schwadel & Stout, 2012).
In contrast, it is understood that if the decline in social capital manifests
because individuals change their minds then there are period effects at work.
While much of the research examines change over time, few consider the
possibility of both period and cohort changes in social capital by simultane-
ously modeling age,2 period, and cohort (APC) effects on social capital
(Schwadel & Stout, 2012). Dramatic changes in social capital levels may
occur as a result of either or both and, the specific source of change can have
important—albeit different and possibly, negative—consequences on soci-
ety. The theoretical expectation is that social change occurs as a gradual pro-
cess as a result of cohort replacement or as a major shift as a consequence of
individual attitude change reflecting the impact of events and movements in
the external world (i.e., time-period effects). The specific source of social
capital trends provides a guide for the potential to reduce or even reverse the
downward trend. Generational effects, unlike period effects, imply that

---

Clark
571
America’s declining social capital will not easily recover. Consequently, it is
essential to distinguish between period and cohort effects when researching
changes in social capital over time.
The problem is compounded by the fact that in the limited number of stud-
ies that have modeled APC effects, the extent to which any such temporal
trends are a consequence of differences in the key predictors of social capital
remains unexplored. The literature suggests that variations in social capital
are associated with a number of individual level and contextual characteris-
tics. As noted by Brehm and Rahn (1997), “social capital is an aggregate
concept that has its basis in individual behavior, attitudes, and predisposi-
tions” (p. 1000). According to Costa and Kahn (2003a), we can consider the
production of social capital to depend both upon individual socioeconomic
and demographic characteristic and upon the characteristics of society. It is
the individual attributes (e.g., education, race, religion, income status, etc.)
that shape one’s decision to participate in community life, but the value of
participating depends upon the extent of others’ participation as well.
A contextual characteristic that receives a large amount of attention in
social capital studies is the degree of income inequality. The general conclu-
sion reached is that the strong relationship between economic conditions and
social capital trends reflects the diminished propensity of people to trust or
spend time socializing with those who are dissimilar from themselves
(Alesina & La Ferrara, 2002). Uslaner and Brown (2005) argue that when
inequality is high, people at the top and the bottom of the income distribution
will not perceive each other as facing a shared fate. Therefore, they will have
less reason to trust or engage with people of different backgrounds.
Furthermore, a number of recent studies have argued that increasing income
inequalities in Western societies intensifies social boundaries (Elgar &
Aitken, 2010; Wilkinson & Pickett, 2009). According to this view, societies
that are highly stratified are more socially segregated; it is therefore less
likely that people from different strata meet each other, resulting in less trust.
In sum, differing levels of social capital are shaped by the presence or absence
of individual characteristics, but it is also an aggregate concept, embodied as
a feature of society with economic, political, and social conditions influenc-
ing the level of social capital.
Despite the consistent association between social capital and individual-
and macro-level factors found in earlier work (Alesina & La Ferrara, 2002;
Brehm & Rahn, 1997; Costa & Kahn, 2003a; Hooghe, 2003; Stolle, 2002),
few studies incorporate these variables in an APC effect accounting. Though
informative, these studies are limited in their ability to estimate cohort- and
period-based changes in social capital as many use statistical techniques that
prevent the inclusion of explanatory variables that can be used to explain the
estimated patterns of these effects.

---

572
American Politics Research 43(4)
Thus, despite the quality of prior studies, the results are restricted to
describing how social capital develops and changes across cohorts and over
time but they are not able to explain why social capital patterns differ from
one generation to the next. Until the association between key individual-level
predictors and aggregate-level change in social capital rates is jointly
assessed, it becomes impossible to quantify the influence of historical hap-
penings or that the effect on cohort has more to do with the formative experi-
ences of the different generations than with compositional (i.e., social and
demographic) differences across cohorts. For instance, Nie, Junn, and
Stehlik-Barry (1996) claim that while education has a positive impact on
civic engagement at the individual level, in aggregate terms the rise in the
average education level in society wipes out this effect. Likewise, a decline
in social capital across period could occur if the relationship between a char-
acteristic (such as education) has changed over time. In short, much remains
unknown about how changes in social capital should in fact be understood.
Here, I expand the literature by taking advantage of a new method of esti-
mation to gauge the degree of variability...