Incentives and gaming in a nonlinear compensation scheme. Evidence from North American auto dealership transaction data

Publication Date07 December 2015
AuthorHideo Owan,Tsuyoshi Tsuru,Katsuhito Uehara
SubjectHR & organizational behaviour,Global HRM
Incentives and gaming in a
nonlinear compensation scheme
Evidence from North American auto
dealership transaction data
Hideo Owan
Institute of Social Science, The University of Tokyo, Tokyo, Japan
Tsuyoshi Tsuru
Institute of Economic Research, Hitotsubashi University, Tokyo, Japan, and
Katsuhito Uehara
Faculty of Human Studies, Tenri University, Nara, Japan
Purpose Under a discontinuous and nonlinear compensation scheme, which is prevalent among car
dealerships, the amount of a salespersons expected daily commission depends primarily on his
position in the pay schedule on the day he makes a sale. Salespeople thus vary their efforts and adopt a
different pricing strategy week by week, or even day by day. The purpose of this paper is to examine
the incentive effect of such a nonlinear scheme and provide the evidence that salespeoples behavior is
consistent with the theory.
Design/methodology/approach The authors conduct regression analyses using the transaction
data provided by two North American auto dealerships. The authors construct a daily measure of
varying incentive intensity and evaluate its impact on the distribution of individual daily sales and the
dealerships gross profit rate.
Findings The authors find that the daily measure of varying incentive intensity has a positive
effect on the distribution of individual daily sales and a negative impact on the dealerships gross
profit rate. The results suggest that: salespeople adjust their effort levels in response to the
intensity of incentives; and they game the system by lowering the prices when the marginal return to
doing so is high.
Research limitations/implications The study shows that there is a high cost associated with the
discontinuous nonlinear pay scheme, raising the question of why many auto dealerships use it.
Originality/value This paper sheds light on the undesirable aspects of discontinuous and
nonlinear incentive schemes, varied performance and gaming, by quantifying the effects of the
workers behavior.
Keywords Personnel economics, Employee motivation, Gaming
Paper type Research paper
1. Introduction
Why are discontinuous, nonlinear pay schemes prevalent in sales operations such as
auto dealerships? From a theoretical point of view, offering nonlinear incentive
contracts is not puzzling. Many early works in contract theory have shown that the
optimal compensation scheme can take a complex nonlinear form, which is sensitive to
the information structure in the model[1]. For example, Basu et al. (1985) show that the
optimal contract could be convex when the absolute coefficient of risk aversion rapidly
decreases with income, and argue that commonly used compensation plans for sales
staff that involve salary, commission, and bonus approximate a convex shape. Mc Afee
and McMillan (1987), on the other hand, show that the optimal pay scheme may be
convex because of the need to sort out capable agents from incapable ones when there
Evidence-based HRM: a Global
Forum for Empirical Scholarship
Vol. 3 No. 3, 2015
pp. 222-243
©Emerald Group Publishing Limited
DOI 10.1108/EBHRM-09-2014-0023
Received 3 September 2014
Revised 14 February 2015
Accepted 27 February 2015
The current issue and full text archive of this journal is available on Emerald Insight at:
is an adverse selection problem. In fact, some of the previous studies on contract theory
focus on the issue of how the use of linear contracts could be justified (e.g. Holmstrom
and Milgrom, 1987).
Nonetheless, discontinuities rarely appear as part of the optimal contract in
tractablemodels under the assumption of risk-averse agents. Although Mirrlees
(1974) and Holmstrom (1979) have shown that two-step (discontinuous) schemes
that punish the agents very severely (albeit with low probability) could approximate
a first-best solution, this type of contract has been criticized as unrealistic for two
reasons. First, the two-step scheme encourages effort by punishing rare disastrous
performance, and thus its effectiveness is very sensitive to the specific assumptio ns of
the probability distribution of such rare events. It is hard to believe that any designer
of incentive pay schemes can accurately estimate such small probabilities. Second, if
the agents can gain private information on their expected performance, they will stop
putting forth the requisite level of effort whenever they become certain that they can
avoid a disaster. Hence, the two-step schemes will lead to unstable variable effort levels,
which would not be an efficient outcome.
In the real world, on the other hand, there have been reports by practitioners and
empiricists that convex or discontinuous pay schemes could have a number of
undesirable side effects.First, such schemes are known to be prone to timing
gamingby employees (see Larkin, 2014 for empirical evidence). Under the convex pa y
scheme, for example, concentrating outputs/sales in one evaluation period increases the
overall pay, thus creating the incentive to manipulate the timing of outputs/sales.
Salespeople might therefore postpone the sales (push out) or advance the sales that
they would otherwise realize in the next period with extra incentives (pull in)
depending on their performance in the early stage of the evaluation period. In addition,
when the pay schedule has discontinuous points, namely, thresholds that trigger
a jump in piece rate or commission rate, the employees who are close to the thresholds
may have an incentive to manipulate the timing of their sales in order to pass the
threshold and earn more.
Second, discontinuities could be disruptive to the business: although employees
close to the thresholds might redouble their efforts, those who have passed the
thresholds might slack off, undermining attempts to induce uniform and stable
effort. Although these issues are quite important for managers, the pros and cons
of discontinuous, nonlinear incentive schemes have not been fully examined in the
economics literature.
This paper attempts to investigate such contractscons gaming and varied
performance and estimate the size of the distortions. We only speculate on their pros,
owing to our belief that proper analysis would require the use of approaches,
such as behavioral science, that cross the traditional boundaries of economics. More
specifically, we answer the following questions: first, how will the changing incentive
intensity along a discontinuous nonlinear pay schedule affect daily productivity; and
second, how will salespersonspositioning in the pay schedule influence the prices they
charge (possibly to advance the timing of customer purchases)?
Our main findings are as follows: first, we find that an incentive intensity measure
that varies over the course of monthly performance is positively associated with daily
sales. This is clear evidence that salespeople put forth either greater or lesser effort
according to the strength of incentives. Second, we find the marginal commission
(the increase in monthly commission income gained by selling each additional car) and
the gross profit rate for each car sale to be negatively correlated after controlling for car
North American
auto dealership
transaction data

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT