Productivity in the Workplace: Cops, Culture, Communication, Cooperation, and Collusion

• DOI: 10.1177/106591290405700106
• Published date: 01 March 2004
• Date: 01 March 2004
• Author: Connie Jorstad,Laura Langbein
• Subject Matter: Articles

Productivity in the Workplace: Cops, Culture,
Communication, Cooperation, and Collusion
LAURA LANGBEIN and CONNIE JORSTAD, AMERICAN UNIVERSITY
Both formal theory and experimental evidence have shown that repeated interactions among actors foster
norms of trust and cooperation. But real-world empirical evidence regarding the substantive effects of repeated
interaction is scant and fails to disentangle the components of “social capital.” Using data from a 1994 survey
of urban police in three Caribbean countries, clustered by police stations, we regress measures of time spent
on police work and of propensity to use excessive force on individual and aggregate-level measures of norms
and communication and a size-related measure of opportunities for repeated interaction. We control for mon-
itoring by superiors, task, experience, gender and other variables. We anticipate non-independent observations
within police stations and heteroscedasticity between as expectations from the theory of social capital rather
than simple statistical nuisances. Face-to-face communication at the individual and police-station level, norms
at the individual and station level, and opportunities for repeated interaction, contribute to increased work
effort by individual officers, and to reduction in their propensity to use excessive force. Monitoring by superi-
ors has no impact on productivity, and may actually have an adverse effect on excessive force. “Social capital”
is simultaneously an individual and aggregate phenomenon. In the presence of pro-social norms, social capi-
tal “trumps” monitoring as an effective tool of management. Fostering social capital among peers may reduce
monitoring and other transactions costs within organizations, making them relatively more cost effective.
The purpose of this article is to use the idea of social people together to cooperate and work reciprocally, for
capital, conceptualized as a repeated game about
social “good” or “bad.”
norms and behavior, as part of an explanation for
The article also adds a note to quantitative comparative
why some individuals in public bureaus, and hence some
studies by examining bureaus in three different Caribbean
bureaus, are more (or less) “productive” than others.1 This
nations, revealing that differences within these nations are
research adds specificity and non-experimental evidence to
sufficient to explain the public productivity differences
the considerable but disjointed literature on social capital. It
between them. Finally, it applies the theory in the context of
separately measures both individual and contextual charac-
the police, a typical public bureau whose output is jointly
teristics, and identifies opportunities for repeated interac-
produced, hard to measure, unrelated to pay, and some-
tions that characterize “social capital.” As part of the measure -
times unwanted. Simple theories of self-interest usually pre-
ment of social capital, this study regards non-independent
dict that rational police officers should do as little as possi-
observations within a workplace and heterogeneity between
ble, no matter whether their efforts promote social “good” or
workplaces not as statistical nuisances to be eliminated but
“bad.” Yet in fact we observe variations in desired and unde-
rather as phenomena to be expected. These “sticky” statisti-
sired effort among police within and between police sta-
cal issues are also the “glue” of social capital that binds
tions, and between nations. This paper shows that social
capital, properly measured, can help to explain these
observed variations in a real world setting.
1 Long ago, Chester Barnard (1938), and many others since, noted the
importance of social relationships for the smooth operation of work-
BACKGROUND
forces. However, these early observational studies did not even begin to
separate cause from effect, failed to hold confounding variables constant,
did not specific exactly what aspects of social relations were important,
One possible explanation for variance in work habits
and did not separate individual from group properties. Miller (1992:
within the public sector may lie in the relatively recent “dis-
179-226) mentions many of the early studies. This study, and many of
covery” among rational choice scholars that people are not
those cited in this study, adds formality and specificity to the theory,
“rational” in the simple way predicted by an unrepeated PD
measurement and testing of early observations about the importance of
game. They vote, join advocacy groups, sometimes con-
social relations in the workforce. Ichniowski and Shaw (2003) report
similar evidence about the impact of “connective capital” for sharing
tribute to public TV or radio, and often cooperate with
information and communicating norms on productivity in the private
others in work settings, choosing to “work” while refusing
sector, especially when tasks are complex.
the dominant “defect” or “shirk” option (Moe 1980; Kropf
NOTE: We thank Dr. Richard R. Bennett for graciously allowing us to use
1998; Brunner 1998; Isaac, Walker, and Thomas 1984;
his data. We also thank the anonymous reviewers for their helpful
Ostrom 1998; Brehm and Gates 1997; Miller 2000; Miller
and insightful suggestions, and for the care with which they read
and Whitford 2002; Miller 1992).
this manuscript. We claim the remaining errors.
The multiply birthed idea of “social capital” lies at the ro o t
Political Research Quarterly, Vol. 57, No. 1 (March 2004): pp. 65-79
of most explanations of cooperative behaviors such as these
65

---

66
POLITICAL RESEARCH QUARTERLY
(Putnam 2000: 446). Putnam points out that social capital
of social capital. When police trust one another to do what
p e rtains not just to characteristics of institutions that cre a t e
is socially desirable (the positive face), that is usually
incentives (Coleman 1988: S98) and disincentives (Loury
described as “cooperation.” When police trust one another
1977) for people to cooperate. It also pertains to re l a t i o n s
to do what is socially undesirable (the dark side), that is
between individuals, particularly the relationship of tru s t
usually described as collusion. This study examines both
and re c i p rocity (Putnam 2000: 20). However, with respect to
faces of social capital by looking not only at the amount of
the workplace (and elsewhere), trust alone has been key to
communication among officers as a proxy for ability of offi-
many explanations of cooperative behavior, perhaps because
cers to predict each others behavior (trust), but also at the
it appears relatively easy to model and measure (Wi n t ro b e
prevailing norms of behavior.
and Breton 1986; Hardin 2000; Schotter 1998; Ostro m
Norms and communication capture much, but not all, of
1998; Miller 2000, 2001; Uslaner 2000, 2001; Brehm and
the amorphous concept of social capital. Repeated, bilateral
Gates 2001; Chen 2000). But, as we will see below, there is
interaction is also critical to the idea of social capital, espe-
m o re to the idea of social capital than trust.
cially in small groups. Orbell, van de Kraght, and Dawes
Trust is usually described as the degree of confidence that
1991; Frank 1991; Ostrom 1998; and Gibbons 2001 show
one person has that another won’t cheat, will deliver on a
that a network of repeated bilateral communications
contract, or will behave predictably (Wi n t robe and Bre t o n
increases cooperation rates, as does smaller group size. They
1986; Hardin 2000; Claibourne and Martin 2000; Bre h m
further argue that this happens not because discussion
and Gates 2001; Tyler 2001). Wi n t robe and Breton (1986)
reveals the costs of non-cooperation, but rather because
f u rther distinguish between vertical and horizontal tru s t ,
when cooperative acts are observed and there is the possi-
contending that vertical trust between employees (agents)
bility of repeated interactions, acts at one time create expec-
and managers (principals) increases workplace pro d u c t i v i t y,
tations about the future. People cooperate if they expect
while horizontal trust among employees diminishes it. How-
everyone else does also; numerous and repeated bilateral
e v e r, this argument illustrates a primary problem with using
communications create and reinforce this expectation. Espe-
the concept of trust alone to explain cooperative behavior, at
cially when workforces are relatively stable, repetition also
least in the workplace. Trust is not simply repeated contacts
solves the last-period problem and keeps promises about
and predictable behavior. In the case of vertical trust, Wi n-
future cooperation from unraveling, since no one knows
t robe and Breton implicitly assume that boss/principal P pre-
when the last period will be.2
dicts that employee/agent A will do what P wants, and that
Thus, many re s e a rchers add the concept of opport u n i t i e s
Agent A predicts that if he does what P wants he, A, will be
for repeated contacts, facilitated by face-to-face communica-
re w a rded. Their further assumption is that horizontal tru s t
tion, to the recipe for social capital (Miller and Whitford
means that employees/agents A1 and A2 will trust each other
2002; Scholz 1991; Schneider, Teske, and Martin 1991;
to shirk, collude, or in some way evade doing what P wants.
O s t rom and Ahn 2001; Ostrom 2000). In arguing for the
Hidden in the argument about trust is the role of norm s :
i m p o rtance of repetition and communication, Miller and
Wi n t robe...