This is an
author-produced blog post designed to introduce upcoming Urban Affairs
Review articles. This article is now available in OnlineFirst.
*****
The
Relationship between Population Size and Contracting Out Public Services:
Evidence from a Quasi-experiment in Danish Municipalities[i]
Department
of Political Science
University
of Copenhagen
E-mail: skf@ifs.ku.dk
1. The
relationship between local government size and contracting out is negative for
capital-intensive services, and is positive for labor-intensive services that
are difficult to measure and/or are regulated by a free choice market.
What is the causal relationship
between population size and the contracting out of public service delivery in
local governments? Robust results from this article show that the
relationship differs across policy sectors: It is negative for services with
high fixed costs, presumably due to scale economies, and is positive for
services that are difficult to measure, probably due to more administrative and
technical capacity in larger municipalities. Also, the effect of population
size is positive for tasks in free-choice markets, presumably because private
contractors find large free-choice markets more attractive.
2. The
results make sense of existing research which have found mixed evidence for the
relationship between local government size and contracting out
Results of the
population size/contracting out relationship have been mixed. A survey of 21
studies conducted from 1986 to 2006 reported that nine studies found a positive
relationship, five studies found a negative relationship, and seven studies
reported mixed or insignificant. Since 2006, four newer studies found a
positive association, while four other studies reported no or mixed results.
Analysis of the
literature identifies two shortcomings in estimating the causal relationship
between population size and contracting out. First, all studies measure the
relationship by non-experimental cross sectional research designs, which often
make the estimates vulnerable to threats to internal validity from unobserved
variables and reverse causality. Second, only modest attention has been paid to
sectoral explanations where the relationship between size and contracting out
depends on the service in question. This article addresses these two
shortcomings by using the unique variation created by Denmark’s 2007 local
government structural reform and by examining the relationship between
population size and contracting out across nine different policy sectors.
3. The
results can be generalized to local governments outside Denmark
The results are
primarily valid for the Danish context and for a population increase from an
average of 21,960 inhabitants before reform to 52,387 inhabitants after reform
and with a control group with an average of 46,475 citizens. However, the
findings can be generalized to local governments in other countries that
produce similar services and operate within a comparable institutional
framework. In fact, the reported negative relationship between population size
and contracting out for services with high fixed costs supports the finding in
a survey of 32 international studies that concluded that “studies that analyze
just one service find more evidence that scale economies are a major [negative]
source of privatization” (Bel and Fageda 2009)
4. The
findings are important for policy-makers that wish to change local government
size since the effect for contracting out depends on the produced
services
The relationship between
population size and contracting out in local governments is an important
research topic since population size is thought to be a driver behind private
contracting which, in turn, can have various implications for the cost and
quality of publicly financed services. A main argument in favor of contracting
out is that private suppliers expose public organizations to competition that
can result in potential cost savings. However, contracting out can also lead to
reduced service quality or be viewed as a step towards more fundamental
privatization of public services (Blomquist 2004). Moreover, because
population size presumably affects the production incentives of both
municipalities and private suppliers, public authorities may wish to influence
the level of contracting out, for example, by manipulating jurisdiction size.
5. The
findings show the conditional nature of the factors influencing privatization
decisions and encourage more research in the politics behind privatization
Besides local government
size, the literature generally identifies three main drivers behind contracting
out: fiscal stress, public preferences, and the power of interest groups. With
respect to this broader literature, this article contributes to recent advances
in the privatization field in determining the causal effects of the drivers
behind contracting out. Furthermore, it contributes specifically to a growing
academic interest in determining the conditional nature of the factors
influencing privatization decisions by accounting for the conditional effect of
population size due to different service and market characteristics. In a more
general sense, it is likely that other drivers of contracting out also work
differently depending on context-specific characteristics.
Such characteristics
should probably not be confined to market and service features, but could be
broadened to include the strength of interest groups within various policy
sectors. For instance, the political clout of public employees and service
users has been used to explain variations in the level of contracting out
across policy sectors in Danish municipalities. Following this logic, the
literature could start to further unpack the conditional effect of interest
groups and for instance investigate to what extent market characteristics can
be explained by the strength of various interest groups that prefer particular
market forms.
[i] Acknowledgments: The authors would like to acknowledge the helpful comments
received from Yosef Bhatti, Ole H. Petersen and Jacob G. Hariri as well as very
constructive guidance received from the editors and five anonymous reviewers.
