Reputation versus information: The delegation policy when the principal has reputational concerns

• Author: Tsung‐Sheng Tsai,Yasunari Tamada
• DOI: http://doi.org/10.1111/jpet.12289
• Published date: 01 June 2018
• Date: 01 June 2018

Received: 8 August 2017 Accepted: 13 January 2018

DOI: 10.1111/jpet.12289

ARTICLE

Reputation versus information: The delegation

policy when the principal has reputational

concerns

Yasunari Tamada1Tsung-Sheng Tsai2

1KeioUniversity

2NationalTaiwan University

YasunariTamada,Faculty of Economics, Keio

University,2-15-45 Mita, Minato-ku, Tokyo

108-8345,Japan (tamada@econ.keio.ac.jp).

Tsung-ShengTsai,Department of Eco-

nomics,National TaiwanUniversity, No.

1,Sec. 4, Roosevelt Road, Taipei106,

Taiwan(tstsai@ntu.edu.tw).

Wesincerely thank the Associate Editor and

tworeferees for their insightful comments and

helpfulsuggestions. The second author also

acknowledgesthe financial support provided by

theMinistry of Science and Technology,Taiwan

(NSC93-2415-H-001-033).

We analyze the delegation policy when the principal has reputa-

tional concerns. Both the principal and the agent can be either good

or biased; the good players prefer the correct decision, while the

biased ones prefer a high action even though it may be wrong. An

evaluator who forms the principal's reputation may or may not be

able to observe who makes the decision. When the evaluator can-

notobserve the allocation of authority, the principal shares the credit

and blame with the agent. Although delegation can improvethe qual-

ity of decision making because the agent has better information, it

mayalso hurt the good principal's reputation because the high action

may be taken by the biased agent while the blame will be shared

when it is wrong. Thus, the good principal has a tendency to keep

too much authority to maintain her reputation. By contrast, when

the evaluator can observe the allocation of authority, delegation

becomes a signaling device for the good type of principal to differ-

entiate herself from the biased one. This results in an excessiveallo-

cation of authority toward the agent.

“Princes should delegate to others the enactment of unpopular measures and keep in their own hands the means of winning

favors.” — Niccolò Machiavelli

1INTRODUCTION

The epigraph noted by Niccolò Machiavelli 500 years ago underlines a fundamental and important question faced by

every decision maker:when should the decision-making authority be delegated to others? In reality, the degree of del-

egation of authority has not been monolithic but has varied in different issues and areas. For example, Epstein and

O'Halloran (1999) demonstrate that the U.S. Congress delegates the least authority to the Executivein issues related

to budgets, rules, and ways and means, but delegates the most in the areas of agricultural and public works and armed

services. Graham, Harvey,and Puri (2015) find that CEOs delegate less when they are more knowledgeable but dele-

gate more decisions when they are overloaded or when they are distracted byrecent acquisitions. Researchers have

Journal of Public Economic Theory.2018;20:367–389. wileyonlinelibrary.com/journal/jpet c

2018 Wiley Periodicals,Inc. 367

368 TAMADAAND TSAI

been providing some good reasons for the principal to delegate the authority to the agent. One important reason is

to take advantage of the agent's expertise. If the principal delegates the authority,the quality of information may be

better, so that it is more likely that the right decision can be made; however, the difference in their preferences may

also lead to a biased decision. The principal thus faces a trade-off between a loss of control and a loss of information in

determining her delegation policy.

Another popular reason for delegation that has been emphasized, particularly in the political science literature,

is blame avoidance. Politicians who seek reelection are often motivated by the desire to avoid blame for unpopular

actions that might impair their reputation or the chance to win in the election. Thus, politicians tend to adopt some

strategies to avoid blame, such as agenda limitation, scapegoating, or “passing the buck.” Accordingly, delegation is

the result of politicians'efforts to “shift the responsibility” (Fiorina, 1982). For example, Congress often passes pro-

tectionist trade legislation, in the hope that the President will veto it and bear the wrath of the affected industries.

Independent regulatory commissions are often delegated responsibility for missions with conflicts that benefit some

industries at the expense of others, such as mergers or rate-making.1On the other hand, politicians also often seek to

“claim credit” through their activities for popular actions (Fiorina, 1977). Such an incentive encourages the principals

to make the decisions on their own to protect their political benefits.2In other words, policymakers care about their

reputation and may choose the actions that help them avoidblame or claim credit rather than the correct ones.

So far,these two points of view in analyzing the delegation policy have been considered separately in the literature.

This paper incorporates both perspectives to determine the delegation policy by introducing a new factor into the

typical trade-off between information and control: the principal's reputational concerns. Consider the situation where

there are three players: a principal (she), an agent (he), and an outside evaluator (he).One can interpret the principal

as a politician or a CEO, the agent as a bureaucrat or a division manager,and the evaluator as the median voter or the

stakeholder.Both the principal and the agent can be either good or biased, which is private information to themselves.

The good type prefers to choose a correct policy, and the biased type prefers some specific policy with a high action.

For example,if a politician plans to build up a public good, the good type will prefer a suitable size which can meet the

voter's demand, while the biased type will always want to build the largest size, which may be considered a spendthrift's

behavior.The principal can either make the decision on her own or delegate it to the agent, who has better information

regarding the state of the world. The evaluator prefers a correct decision to a wrong one. He also forms the principal's

reputation after observing the decision that has been made. Importantly,the principal cares not only about the chosen

decision but also her reputation. The interesting feature of our analysis is that the principal can use her delegation

policy to manipulate her reputation.

The biased principal's favorite policy is the high action, and therefore, when reputational concerns are not too

strong, she prefers to keep the authority all the time because the low (although correct) action may be chosen by the

good agent under delegation. An interesting question follows: Towhat extent would the good principal delegate her

authority, given that the biased type always holds on to power? Besides facing the typical trade-offbetween control

and information, her reputational concerns also play an active role. Importantly,the formation of her reputation cru-

cially depends on whether the allocation of authority can be observed by the evaluator or not. When the evaluator

cannot observe the allocation of authority, the responsibility for the decision making is ambiguous and the blame or

credit will be shared between the principal and the agent. Therefore, if the authority is delegated, the good principal

facesa new trade-off between utilizing the agent's better information and maintaining her reputation. On the one hand,

as the agent is more likely to take the correct action and only errs in the high action, the quality of decision making is

better.However, it is also likely that the high action is taken by the biased agent, which is also preferred by the biased

principal. As the blame is shared, delegation that leads to the high (and incorrect) action can hurt the good principal's

reputation.

1SeeEpstein and O'Halloran (1999) and Weaver (1986) for the detailed discussions.

2Forexample, Epstein and O'Halloran (1999) state that Congress uses considerable resources to write detailed legislation about tax policy and leaves little

authorityto the executive branch. The advantages of controlling tax policy come from the possible benefits of grantingcorporations and interest groups special

taxbreaks in exchange for their support or contributions in elections.