Principal-Agent Modelling and Learning
| Published date | 01 October 2007 |
| Date | 01 October 2007 |
| DOI | http://doi.org/10.1177/0952076707081585 |
| Subject Matter | Articles |
Principal-Agent Modelling and
Learning
The European Commission, Experts and Agricultural
Hormone Growth Promoters
Claire A. Dunlop and Oliver James
University of Exeter, UK
Abstract Principal-agent modelling has become a very influential way of thinking
about bureaucratic politics in a wide range of settings. Simple agency models
have recently been extended and bureaucratic relationships placed in their
wider temporal and socio-political contexts. By developing a
conceptualization of principal-agent learning this article offers a nuanced
account of the extent to which principals can learn to develop institutions that
enhance their political control of bureaucratic agents and predispose those
agents toward the principal’s preferences, limiting adverse selection and
moral hazard problems. The revised model is applied to the context of the
European Commission’s selection and management of scientific committees
in the case of agricultural hormone growth promoters. The findings not only
confirm the usefulness of more dynamic accounts of principal-agent
relationships that eschew ahistoricism and acontextualism, they also suggest
that extended principal-agent models should include constraints on learning
by principals. In this case the principal’s learning was reactive – relying
heavily upon external actors and venues to do their thinking for them.
Keywords adverse selection, European Commission policy making, expert advice,
learning, moral hazard, principal-agent
Introduction
Over two decades ago, Terry Moe (1984) accurately predicted principal-agent
modelling would become an integral part of organizational analysis. Since
DOI: 10.1177/0952076707081585
Claire Dunlop, University of Exeter, Department of Politics, Amory Building, Rennes Drive,
Exeter, Devon EX4 4RJ, UK. [email: c.a.dunlop@ex.ac.uk]
Oliver James, University of Exeter, Department of Politics, Amory Building, Rennes Drive,
Exeter, Devon EX4 4RJ, UK. [email: o.james@ex.ac.uk] 403
© Public Policy and Administration
SAGE Publications Ltd
Los Angeles, London, New Delhi
and Singapore
0952-0767
200710 22(4) 403–422
becoming perhaps the single most dominant paradigm for the analysis of bureau-
cratic politics, the approach has attracted criticism from some scholars trying to
use it. Notably, traditional agency analyses’ depiction of bureaucratic relation-
ships as static dyads unaffected by externalities thrown-up by the socio-political
world within which they are situated has come under attack (Mitnick, 1992). In
response, the ‘extended principal-agent’ approach has been suggested where
bureaucratic relationships are placed in their wider temporal and socio-political
contexts and information asymmetries and goal conflicts treated as dynamic rather
than fixed variables. By doing this, extended analysis opens up the possibility that
interaction with additional actors and venues may change the patterns of informa-
tion available to principals and agents, thereby transforming how they relate to one
another (Meier and Krause, 2003; Waterman and Meier, 1998).
Extended analysis is built upon the assumption that principal-agent relation-
ships evolve; as externalities change the supply of information received by
principals and agents over time, the capacity of both sets of actors to learn about
policy and politics will increase. Various scholars have explored this thesis from
the viewpoint of agents (notably, Waterman and Rouse, 1999; Waterman et al.,
1998). However, less attention has been paid to the ways in which principals learn
from additional actors and venues.1The implications of learning are pronounced
for principals – any loss of control to agents need not be as inevitable or irrevoca-
ble as the traditional agency model suggests. How principals learn and the degree
of control they have over the learning process represent the main foci of the
article. What type of control can principals exert over learning? Specifically, to
what extent can they engage proactively with agents and their environment
over time to select and manage agents in ways that benefit their own policy
preferences?
Using a detailed conceptualization of different sources and types of learning the
article explores how, through interaction with additional actors, principals can
learn to overcome three collective action dilemmas familiar to agency analysis.
The revised model is applied to the case of the European Commission’s selection
and management of scientific committees in the case of agricultural hormone
growth promoters.
The research presented in this article breaks with the methodological norm of
agency analysis. Primary interviews and documentary analysis are used here to
track the ways in which principals learned about the issue and the expert agents
they could assemble to support their policy stance. The article’s use of qualitative
evidence speaks directly to Moe’s (1984: 758) assertion that agency analyses’
enduring contribution is conditional upon the incorporation of a ‘more eclectic
methodology’ where the parsimony of agency theory could be usefully balanced
by organizational approaches more concerned with empirical richness. Scholars
are increasingly aware of the need ‘to adopt methods that correspond to the data-
generating process in bureaucracies’ (Meier and Krause, 2003: 305). In exploring
the role of learning in bureaucratic politics the empirical challenge is to understand
Public Policy and Administration 22(4)
404
To continue reading
Request your trial