Balanced Scorecards: A Relational Contract Approach

• Published date: 01 May 2023
• Author: OLA KVALØY,TROND E. OLSEN
• Date: 01 May 2023
• DOI: http://doi.org/10.1111/1475-679X.12465

DOI: 10.1111/1475-679X.12465

Journal of Accounting Research

Vol. 61 No. 2 May 2023

Printed in U.S.A.

Balanced Scorecards: A Relational

Contract Approach

OLA KVALØY∗AND TROND E. OLSEN†

Received 24 February 2021; accepted 3 October 2022

ABSTRACT

Reward systems based on balanced scorecards often connect pay to an in-

dex, that is, a weighted sum of multiple performance measures. We show that

such an index contract may indeed be optimal if performance measures are

nonverif‌iable so that the contracting parties must rely on self-enforcement.

Under commonly invoked assumptions (including normally distributed mea-

surements), we show that the weights in the index ref‌lect a tradeoff between

distortion and precision for the measures. The eff‌iciency of the contract im-

proves with higher precision of the index measure, because this strengthens

incentives, and correlations between measurements may for this reason be

benef‌icial. There is a caveat, however, because the index contract is not nec-

essarily optimal for very precise measurements, although it is shown to be

asymptotically optimal. We also consider hybrid measurements, and show that

the principal may want to include verif‌iable performance measures in the

∗University of Stavanger Business School and Department of Business and Management

Science, Norwegian School of Economics; †Department of Business and Management Science,

Norwegian School of Economics

Accepted by Haresh Sapra. We are grateful for valuable comments and suggestions from

two referees and Jurg Budde, Bob Gibbons, Marta Troya Martinez, and conference and sem-

inar participants at the 3rd Workshop on Relational Contracts at Kellogg School of Manage-

ment, the 30th EALE conference in Lyon, the 11th Workshop on Accounting Research in

Zurich, and the EARIE 2019 conference in Barcelona.

An Online Appendix to this paper can be downloaded at https://www.chicagobooth.edu/jar-

online-supplements.

619

© 2022 The Authors. Journal of Accounting Research published by Wiley Periodicals LLC on behalf of The

Chookaszian Accounting Research Center at the University of Chicago Booth School of Business.

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.

620 o. kvaløy and t. e. olsen

relational index contract in order to improve incentives, and that this has

noteworthy implications for the formal contract.

JEL codes: D86, J33, M52

Keywords: balanced scorecards; performance measures; relational con-

tracts

1. Introduction

Very few jobs can be measured along one single dimension; employees usu-

ally multitask. This creates challenges for incentive providers: If the f‌irm

only rewards a subset of dimensions or tasks, agents will have incentives to

exert efforts only on those tasks that are rewarded, and ignore others. A

solution for the f‌irm is to add more metrics to the compensation scheme,

but this usually implies some form of measurement problem, leading ei-

ther to more noise or distortions, or to the use of nonverif‌iable (subjective)

performance measures.

The latter is often implemented by the use of a balanced scorecard

(BSC). Kaplan and Norton’s [1992, 1996] highly inf‌luential concept began

with a premise that exclusive reliance on verif‌iable f‌inancial performance

measures was not suff‌icient, as it could distort behavior and promote ef-

fort that is not compatible with long-term value creation. Their main ideas

were indebted to the canonical multitasking models of Holmström and Mil-

grom [1991] and Baker [1992]. However, their approach was more prac-

tical, guiding f‌irms in how to design performance measurement systems

that focus not only on short-term f‌inancial objectives, but also on long-term

strategic goals (Kaplan and Norton [2001]).

While measuring performance is one issue, the question of how to

reward performance is a different one. As noted by Budde [2007], there

is a general understanding that eff‌icient incentives must be based on

multiple performance measures, including nonverif‌iable ones. Still, the

implementation is a matter of controversy. Reward systems based on BSC

often connect pay to an index, that is, a weighted sum of multiple perfor-

mance measures, but there is also variety regarding to what extent index

contracts are used (see WorldatWork and Deloitte Consulting LLP [2014],

and section 5 later in this paper for a discussion). There is apparently no

formal incentive model that actually derives this kind of index contract

as an optimal solution in settings with nonverif‌iable measures.1In fact,

Kaplan and Norton [1996] were sceptical to compensation formulas that

calculated incentive compensation directly via a sum of weighted metrics.

Rather, they proposed to establish different bonuses for a whole set of

1Banker and Datar [1989] derive conditions under which a contract based on a linear

aggregate of verif‌iable performance measures is optimal in a standard moral hazard problem

with a risk-averse agent.

balanced scorecards 621

critical performance measures, more in line with the original ideas of

Holmström and Milgrom [1991] and Feltham and Xie [1994].

Despite the large literature following the introduction of BSC (see

Hogue [2014], for a review), and the massive use of scorecards in prac-

tice, it appears that the index contracts that BSC f‌irms often prescribe lack

a formal contract theoretic justif‌ication.2We take some steps to f‌ill the

gap. Our starting point is that the performance measures are nonverif‌i-

able. This means that the incentive contract cannot be enforced by a third

party and thus needs to be self-enforcing—or what is commonly termed

“relational.” Incentive contracts used by f‌irms, including performance mea-

sures based on BSCs, often include nonverif‌iable qualitative assessments of

performance (see Ittner, Larcker, and Meyer [2003], Gibbs et al. [2004],

Kaplan and Gibbons [2015]). Moreover, even if some performance mea-

sures in principle are verif‌iable, the costs and uncertainty of taking the

contract to court may be so high that the parties in practice need to rely

on self-enforcement (see MacLeod [2007] and references therein).

In the now large literature on self-enforcing relational contracts, rela-

tively few papers have considered relational contracts with multitasking

agents (prominent papers include Baker, Gibbons, and Murphy [2002],

Budde [2007], Schottner [2008], Mukherjee and Vasconcelos [2011], Ishi-

hara [2016], Ishihara [2020]). We on the one hand generalize this liter-

ature in some dimensions (to an arbitrary number of tasks with stochas-

tic measurements that are possibly correlated and/or distorted), and on

the other hand invoke assumptions (notably normally distributed measure-

ments) that make the model quite tractable.3

We f‌irst show that the optimal relational contract between a principal

and a multitasking agent turns out to be an index contract, or what one

may call a BSC. That is, the agent gets a bonus if a weighted sum of per-

formance outcomes on the various tasks (an index) exceeds a hurdle. This

is in contrast to the optimal contract in, for example, Holmström and Mil-

grom [1991], where the agent gets a bonus on each task. The important

difference from Holmström and Milgrom is that we consider a relational

contracting setting where the size of the bonus is limited by the principal’s

temptation to renege (rather than risk considerations). In such a setting,

the marginal incentives to exert effort on each task are higher with index

contracts than with bonuses awarded on each task.

2According to Hogue [2014], among the more than 100 papers published on BSC theory,

only a handful have used principal agent theory to analyze BSC. See also Hesford et al. [2007]

for a review.

3Our paper is indebted to the seminal literature on relational contracts. The concept of

relational contracts was f‌irst def‌ined and explored by legal scholars (Macaulay [1963], Macneil

[1978]), whereas the formal literature started with Klein and Leff‌ler [1981]. MacLeod and

Malcomson [1989] provide a general treatment of the symmetric information case, whereas

Levin [2003] generalizes the case of asymmetric information. The relevance of the relational

contract approach to management accounting and performance measurement is discussed in

Glover [2012] and Baldenius, Glover, and Xue [2016].