When the light shines too much: Rational inattention and pandering

• DOI: http://doi.org/10.1111/jpet.12402
• Date: 01 February 2020
• Published date: 01 February 2020
• Author: Federico Trombetta

© 2019 The Authors. Journal of Public Economic Theory published by Wiley Periodicals, Inc.

J Public Econ Theory. 2020;22:98–145.98

|

wileyonlinelibrary.com/journal/jpet

Received: 15 August 2018

|

Accepted: 3 September 2019

DOI: 10.1111/jpet.12402

ORIGINAL ARTICLE

When the light shines too much: Rational

inattention and pandering

Federico Trombetta

Department of Economics, The

University of Warwick, Coventry, UK

Correspondence

Federico Trombetta, Department of

Economics, The University of Warwick,

Coventry CV4 7AL, UK.

Email: f.trombetta@warwick.ac.uk

Funding information

Economic and Social Research Council,

Grant/Award Number: DTC –ES/

J500203/1

Abstract

Should voters always pay attention to politics? I explore

the role of endogenous costly attention allocation in

politics, combining insights from the growing literature

on rational inattention with a standard model of political

agency. I show that when attention to the action of the

politician is endogenous, voters may choose to pay too

much attention in equilibrium, and this induces too

much political pandering. Moreover when attention to

the action and to the state of the world are both

endogenous, voters may not pay enough attention the

state with respect to the ex ante optimum. A reduction in

the total cost of attention does not correctthis inefficiency

and can even reduce welfare. This model can be a

demand‐driven explanation of the under‐provision of

analytical contents by news channels.

KEYWORDS

pandering, political agency, populism, rational inattention,

social media

JEL CLASSIFICATION

D72; D78

1

|

INTRODUCTION

Are voters paying attention to the right elements when judging, for example, the political

implications of a wall on the United States–Mexico border? Despite concrete doubts over the

effectiveness of the wall, both in terms of reducing illegal immigration and fighting drug

----------------------------------------------------------------------------

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and

reproduction in any medium, provided the original work is properly cited.

cartels,

1

President Donald Trump seems to insist on this political choice. One possible

explanation is that it strengthens in voters the idea of a President aligned with America’s

interests, something that the voters would notice and appreciate. At the same time, maybe,

those voters are not paying too much attention to the effectiveness of the wall itself.

When 24 hr news channels and social media provide constant access to political news, what

is the role of voters’attention in shaping political decisions? Are voters paying the right amount

of attention to politics? And are they paying attention to the right elements? What are the

consequences of inefficient allocation of attention? At first, it seems obvious that voters should

be motivated to pay as much attention as possible to politics, so that they make better choices

and elect better politicians. Moreover, tools that make attention “cheaper”should have an

unambiguously positive impact. I show that reality is more complex: the decision about how

much attention to pay to politics and how to allocate this attention is nontrivial, with

equilibrium incentives that can motivate voters to pay too much attention to what politicians do

(e.g., building a wall on the United States–Mexico border) and too little attention to what

politicians should do (e.g., given the way illegal immigration and drug dealing work, is a wall

the most effective way to achieve the objective?). This is consistent with the media studies

literature that looks at the increasingly important news channels and finds that “80‐85 per cent

of news stories on the news channels contain no context or analytical content at all.”

2

Voters’attention equilibrium allocation may motivate politicians to do what voters want,

irrespective of what would be more appropriate, and voters may be paying too much attention to

some aspects of the political process and too little to others. Given that these allocations might be

suboptimal, policies intended to improve citizen’s attention to politics could be making things worse.

This is the first paper in political economy to look at the effects of voters’rational inattention on

pandering in a political agency framework. The model builds on the literature on pandering, under

which the politician chooses the action that guarantees his re‐election, rather than the optimal one,

for example, Fox (2007) and Maskin and Tirole (2004).

3

My definition of pandering borrows directly

from Besley (2006, chapter 3, footnote 35): pandering occurs when a politician “fails to follow

information about the optimal action for fear of its electoral consequences.”That is, when the

politician chooses the action that gives him higher re‐election chances over the one that he favours

given the state of the world.

4

In that context, this paper endogenizes voters’level of attention to

politics, following the recent and growing literature on the political consequences of rational

inattention (Matejka & Tabellini, 2017; Prato & Wolton, 2016, 2018). On top of that, it takes into

account that attention to politics is a manifold concept. Therefore, this paper endogenizes both the

overall amount of attention to politics and the way attention is distributed between actions and the

state of the world, studying the interaction between the two.

I find novel results based on the fact that the inability to precommit to a certain level of

attention, together with the direction of the relationship between different types of attention

and the politicians’actions, can create important and unexplored trade‐offs.

The model shows that voters’equilibrium allocation of attention is generically suboptimal.

First of all, when only endogenous attention to the action is considered, I show that it increases

pandering and that the equilibrium level of attention chosen by the voter is higher than the ex

1

See, for example, Driver (2018).

2

Lewis, Cushion, and Thomas (2005, p. 470).

3

The idea that political delegation may imply suboptimal choices is also, for example, in Laussel and Van Long (2019), although with a different focus

(optimality of ex‐ante constraints to politicians’actions).

4

Pandering can be a good theoretical approximation of the basic essence of populism. This interpretation is not new in economics, for example, Frisell (2009).

TROMBETTA

|

99

ante optimal level. This is because, when choosing the equilibrium level of attention, the voters

are concerned about the selection of good politicians for the next period, and observing the

action is useful to that purpose. However, they are not taking into account the fact that higher

attention increases the incentive to pander, and this has a negative effect on both period one

voters’welfare and the ability to discriminate between types of politicians.

I then look at how voters allocate attention between action and the state of the world, showing

that, in equilibrium, we tend to pay too much attention to what politicians are doing and too little to

what they should do. In particular I show that, in equilibrium, there is always under‐attention to the

state of the world and I find sufficient conditions for over‐attention to action.

In this model, when selection is the only thing that matters, knowing the state of the world

on top of the action buys nothing to the voter, in expectation. This is because learning the state

of the world is never going to strictly change the voter’sre‐election strategy, and as a

consequence there is no incentive to invest any costly attention there. In contrast, knowing the

action is useful for the voting choice.

From an ex ante perspective, however, attention to the state of the world can be useful in

reducing pandering, leading to a higher first period utility (I use also the term “policy

welfare”for this part of the voter’s objective function), but the equilibrium choice ignores the

disciplining effect of this type of attention, because the politician’s decision is sunk.

Moreover, I show that, since attention to the state reduces only the “bad pandering”(i.e.,

politicians choosing an action different than the state to be re‐elected) but not the “good

pandering”(i.e., a dissonant politician choosing the voter’s preferred action to be re‐elected),

5

while attention to the action increases both, this model exhibits an interesting

complementarity between the two types of attention, that is ignored in the equilibrium

choice. From an ex ante perspective, more attention to the action makes attention to the state

more desirable: reducing bad pandering is moreimportantwhenpanderingisanissue,that

is, when attention to the action is high.

Attention to the action works in the opposite way: it makes pandering more desirable for

the politician. The whole point of choosing a suboptimal action is the electoral reward,

which only happens if this action is observed. In equilibrium the voter does not take into

account this negative effect

6

and as a consequence she tends to allocate too much attention

to the action.

7

As a result of this inefficient equilibrium allocation politicians have

incentives to pander too much, reducing the voters’welfare. Importantly, an exogenous

reduction in the total cost of attention is likely to make things worse, overall, because it

induces an even higher level of over‐attention to action and rewards more pandering,

without improving the level of attention to the state.

Those results help to explain the relationship between social media and populism.

Interestingly, politicians advocating populist policies are over‐represented on social media;

8

social media usage is correlated with political engagement,

9

political attention and consumption

5

“Good”and “bad”are to be interpreted with respect to policy welfare only. This is different from the total welfare of the voter, that comes from the selection

part as well, that is, from the ability to select a good politician in period 2.

6

While the result of no attention to the state is specific of this model, the equilibrium tendency to under‐evaluate the disciplining effect of attention to the state

and under‐consider the downsides of attention to the action is more general, as shown in Appendix F.

7

This is always the case, at least weakly, when the voter has to allocate a certain amount of attention between action and state, as in Appendix D. When the level

of the two attentions is endogenous and the cost function is additively separable there is always under‐attention to the state. Over‐attention to the action

depends on the strength of the ex ante complementarity between the two.

8

Extreme Tweeting, The Economist, November 19, 2015.

9

Wihbey (2015) and Barberá (2018) for a review, although evidence is somehow mixed.

100

|

TROMBETTA