Negative feedback, political attention, and public policy
| Author | Søren K. Larsen,Martin Baekgaard,Peter B. Mortensen |
| Published date | 01 March 2019 |
| Date | 01 March 2019 |
| DOI | http://doi.org/10.1111/padm.12569 |
ORIGINAL ARTICLE
Negative feedback, political attention, and public
policy
Martin Baekgaard | Søren K. Larsen | Peter B. Mortensen
Department of Political Science, Aarhus
University, Aarhus, Denmark
Correspondence
Peter B. Mortensen, Department of Political
Science, Aarhus University, Bartholins Allé 7,
Aarhus 8000, Denmark.
Email: peter@ps.au.dk
Funding information
Samfund og Erhverv, Det Frie Forskningsråd,
Grant/Award Number: 1327-00091
More than 50 years of policy research has provided evidence of
negative feedback where self-correcting mechanisms reinforce sta-
bility in public policies over time. While such mechanisms are at the
heart of understanding change and stability in public policies, little
attention has been given to the responses of individual policy-
makers to public policies as a potential driver of negative feedback.
Based on a unique survey dataset of spending preferences of local
government politicians covering more than 90 Danish municipali-
ties, three years, seven policy issues, and around 3,000 entries, we
find that the expressed spending preferences of politicians are
indeed negatively affected by previous spending levels. Moreover,
such negative feedback effects are stronger, the less the political
attention to the issue and even disappear at high levels of atten-
tion. Our analysis thus provides important evidence on the micro
foundations and conditions of negative feedback in public policy.
1|INTRODUCTION
Negative feedback is a strong force in public policy. It refers to self-correcting mechanisms that reinforce stability.
Just as a thermostat adjusts to falling temperatures by putting out more heat, negative feedback in public policy
maintains stability. Explicit or implicit, most theories and models of public policy include such homeostatic processes.
This includes dominant theories of public policy such as Hall’s (1993) theory of social learning and policy paradigms,
the advocacy coalition framework proposed by Sabatier (1987, 1988), and the punctuated equilibrium theory devel-
oped by Baumgartner and Jones (1993). While these three approaches differ in their conceptualization, description,
and explanation of negative feedback mechanisms, they all depict a policy-making process dominated by negative
feedback. Also the literature on bounded rationality and incrementalism depicts a policy-making process character-
ized by negative feedback (Lindblom 1959; Wildavsky 1964; Simon 1997). Negative feedback is the world of incre-
mentalism where policy action is kept within narrow bounds because most changes to a policy output at time t are
countered by subsequent changes in the opposite direction at time t + 1 (see also Robinson and Meier 2006).
We address two central questions of negative policy feedback, which have not yet been systematically exam-
ined. The first question regards when policy at time t loses its predictive power of policy at time t + 1. Already Wild-
avsky and colleagues were occupied with variants of this question (e.g., Davis et al. 1974) and the question is also
Received: 21 December 2017 Revised and accepted: 2 November 2018
DOI: 10.1111/padm.12569
210 © 2018 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/padm Public Administration. 2019;97:210–225.
prominent in the three more recent approaches to public policy dynamics mentioned above. In particular, we investi-
gate a core claim by Jones and Baumgartner (2005, p. 329): ‘Attention at the individual and collective levels governs
the shift from stasis to the positive feedback processes …’. The quote reflects their emphasis of both negative and
positive feedback, the latter describing instances where an adjustment from the status quo is accentuated by self-
reinforcing processes rather than counterbalanced (Baumgartner and Jones 2002, p. 13). From that perspective, polit-
ical attention—often conceptualized as policy agenda setting—is the central variable governing the mode of public
policy. According to Baumgartner and Jones, negative feedback is the rule when an issue receives relatively low polit-
ical attention, but when the issue attracts political attention the negative feedback mechanism may be weakened or
perhaps even replaced with positive feedback (see also Jones 2001, p. 144). While some case studies have provided
examples in support of this claim (see e.g., Baumgartner and Jones 1993, 2002), the conditional effect of political
attention on negative policy feedback has never been modelled and examined systematically in large-scale empirical
studies. To specify and test this proposition is our first contribution.
The second contribution regards the question of why, most of the time, there will be a thermostatic relationship
between policy at time t and policy at time t + 1. A whole array of explanations of this can be identified in the policy
literature, including variants of institutional friction and/or opposition from vested interests that benefit from the
current policies (e.g., Wilson 1980; Tsebelis 2002). All of these explanations may have merits, but the multiplicity of
explanations is also evidence of a need to begin to disentangle them. The most obvious place to begin, in our view, is
to examine whether negative feedback can be established at the level of individual policy-makers. One of the clear
strengths of many public policy theories is the explicit micro-foundation. Yet, surprisingly little empirical attention
has been given to individual policy-makers. Hence, one likely driver of negative feedback in public policy is that indi-
vidual policy-makers respond to public policies in a thermostatic way very similar to how the public has been shown
to respond to public policies (see Soroka and Wlezien 2010). Politicians can be expected to have knowledge of cur-
rent policies, and if, for instance, spending in an area has increased politicians may respond by lowering their prefer-
ences for spending in that area. Similarly, if spending has been decreasing in a given area, politicians may observe this
and raise their demands for spending in that area. In that way, politicians can be expected to behave in a way similar
to the most ‘informed’part of the public, which according to Soroka and Wlezien’s (2010) studies are the most
responsive to public policies.
Empirically, we address these questions as follows. First, utilizing a unique survey dataset of spending prefer-
ences of local government politicians covering more than 90 Danish municipalities, three years, seven policy issues,
and around 3,000 entries, we show how the expressed spending preferences of politicians generally respond in a
thermostatic, negative feedback style to previous spending levels. This is important individual-level evidence of the
negative feedback mechanism postulated by the punctuated equilibrium theory as well as by other models of public
policy and public spending.
Second, combining the survey and spending data with a dataset on local government policy agendas, we show
how the negative feedback effect disappears as the issue rises on the local political agenda. In other words, as politi-
cal attention to an issue increases, the effect of past spending on current spending preferences wanes. This is sup-
port for the claim that political attention governs the mode of policy, but it also raises new questions about the role
and conceptualization of positive feedback in public policy. The large-Nlocal government data allow us to control for
a wide range of potentially confounding variables, which corroborates the significance of the findings of main theo-
retical interest. The broader implications of our findings, including questions of generalizability, are discussed in the
concluding part of the article.
2|NEGATIVE FEEDBACK IN PUBLIC POLICY
Normally, politics is not explosive. The best predictor of this year’s policy is last year’s policy. Lindblom (1959) and
Wildavsky (1964) were among the first scholars to argue and show this, and even though the quantitative findings, in
BAEKGAARD ET AL.211
To continue reading
Request your trial