Relational contracting and endogenous formation of teamwork

• Date: 01 May 2017
• Author: Akifumi Ishihara
• DOI: http://doi.org/10.1111/1756-2171.12178
• Published date: 01 May 2017

RAND Journal of Economics

Vol.48, No. 2, Summer 2017

pp. 335–357

Relational contracting and endogenous

formation of teamwork

Akifumi Ishihara∗

We study a relational contracting model with two agents where each agent faces multiple tasks:

effort toward the agent’s own project and helping effort toward another agent’s project. We show

that the optimal task structure is either specialization without help or teamworkwith a substantial

amount of help: teamwork with a small amount of help is never optimal. Specialization with high-

powered incentives can be implemented by relative performance evaluation. However, under

teamwork, the evaluation scheme must be substantially different to overcome the multitasking

problem. Consequently, a small amount of help is dominated by specialization with high powered

incentives.

1. Introduction

Recently, organizations have increasingly attempted to encourage cooperative interaction

among workers and to induce teamwork within the workplace (Lazear and Shaw, 2007). Team-

work is potentially valuable when workers can positively influence one another’s productivity by

fostering coordination, information sharing, complementarity of heterogeneous characteristics,

peer monitoring, and psychological peer pressure (Lazear, 1998). However,despite such potential

benefits, team-based organizations frequently exhibit poor performance.1

The implementation of successful teamwork practices is associated with several difficulties

in practice. First, successful teamwork must be combined with appropriate reward schemes for

workers that are consistent with the team objective. Some empirical studies support the premise

that compensation systems significantly influence workers’ incentive to help or compete with

peers in their workplace (Wageman,1995; Drago and Gar vey, 1998; Chan, Li, and Pierce, 2014).

Therefore, measuring performances and rewarding workers are required to establish a successful

∗National Graduate Institute for Policy Studies; a-ishihara@grips.ac.jp.

I am grateful to Kong-Pin Chen, Ricard Gil, Hideshi Itoh, Takeshi Murooka, Hideo Owan, Koji Shirai, Editor David

Martimort, and two anonymous referees for valuable suggestions and the participants of the 6th Japan-TaiwanContract

Theory Conference, 2013 JEA Spring Meeting, 2013 OEIO Summer Conference, EEA-ESEM 2013, The 2014 North

American Meeting of the Econometric Society, EARIE 2014, Kansai Labor Workshop, and seminars at Hitotsubashi,

Kyoto, Nagoya, Osaka Prefecture U, GRIPS, and Academia Sinica for discussion and comments. This research was

supported by financial support from Grant-in-Aid for Young Scientists (B), the Japan Society for the Promotion of

Science, and the Ministry of Education, Culture, Sports, Science, and Technology. All errors are myown.

1Forinstance, Shaw and Schneier (1995) report that although almost all Fortune 1000 companies use project teams,

9 out of 10 teams fail.

C2017, The RAND Corporation. 335

336 / THE RAND JOURNAL OF ECONOMICS

teamwork. Moreover, teamwork management heavily relies on informal rules. It is often impos-

sible to specify working rules by ex ante formal contracting due to difficulties in the description

or verification of state contingencies. Instead, as documented by Gibbons and Henderson (2013),

management practices that achieve high productivity are based on relational contracts in which

informal rules are supported by long-term relationships. As it is easily imagined, teamwork

activity such as dense communication, shared knowledge, peer monitoring, and sophisticated co-

ordination are complex to measure objectively.2Consistent with this view, successful teamwork

is often informally supported by long-term relationships in companies including Toyota (Adler,

Goldoftas, and Levine, 1999) and Southwest Airlines (Gittell, 2000, 2003).3

The aim of this article is to examine teamwork formation through relational contracting.

We build a repeated multitasking principal-agent model based on Itoh (1991), in which there are

multiple agents and each has two-dimensional tasks: (own) effort contributing to his own project

and helping activities contributing to another’s project. In our model, each project generates a

noisy performance measurement, which is observable to both the principal and the agents but not

verifiable. However, in a long-term relationship, a self-enforcing incentive contract based on the

subjective measurement may be feasible.

As an analytical method, we first introduce the two-step approach, which is parallel to the

method adopted in the classic moral hazard problem such as Grossman and Hart (1983) and

Itoh (1991). In relational contracting models, the enforcement problem of informal bonuses is

captured by the dynamic enforcement constraint, which formally states that the sum of the in-

formal bonuses paid to the agents must be less than the future total surplus for any realization

of performance measures. Hence, we first consider the implementation problem in which, given

a targeted decision, the possible maximum amount of the sum of the informal bonuses is min-

imized subject to the agents’ incentive-compatibility constraints. Then, in the second step, the

principal chooses the optimal decision subject to the dynamic enforcement constraint that takes

the incentive-compatibility constraints into account in the first step. This approach is applicable

to other relational contracting models with multiple agents and signals when the optimal decision

level and/or the structure of the bonus structure are of interest in the analysis.

From the two-step approach, we establish the nonconvexity of the organizational pattern

in the sense that the optimal organization must take an extreme form. Specifically, the optimal

organization structure must exhibit either specialization, in which each agent focuses on his own

effort or substantial teamwork, in which each agent chooses a large amount of help. In other

words, intermediate cases in which positive but a small amount of help is provided are never

optimal. The key idea behind this result is that a change in the organization from specialization

to teamwork causes a discontinuous change in the evaluation scheme due to the enforcement

problem of relational contracts.

There is a sharp contrast between specialization and teamwork in the optimal rewarding

scheme, even if the differencein the amount of help is tiny. As the dynamic enforcement constraint

indicates, an informal incentive contract can be enforced if and only if the possible maximum

amount of the sum of the bonuses paid to the agents is below a certain level that depends on

the principal’s patience. Given this enforcement constraint, the agents under specialization are

optimally rewarded according to an extreme form of relative performance evaluation: the agents

compete for some fixed amount of prize and an agent fully receives it if the performance of his

project is better than that of the other agent . However, this evaluation scheme adversely works

under teamwork because an agent prefers the other agent’s performance to be worse and has

an incentive to reduce the amount of help provided. Therefore, as long as the targeted level of

help is positive, to overcome multitasking problems under teamwork, the principal must adopt a

2For instance, Gant, Ichniowski, and Shaw(2002) argue that productive problem-solving activities in a shop floor

are associated with informal networking among workers, which are difficult to be established through formal work

practices.

3Besides, in professional service firms, the business is usually performed by employee teams and the evaluationis

based on subjectively perceivedser vice qualities (Kaiser and Ringlstetter,2011).

C

The RAND Corporation 2017.