Strategic shirking in competitive labor markets: A general model of multi‐task promotion tournaments with employer learning

• Published date: 01 April 2020
• DOI: http://doi.org/10.1111/jems.12342
• Date: 01 April 2020
• Author: Oliver Gürtler,Jed DeVaro

J Econ Manage Strat. 2020;29:335–376. wileyonlinelibrary.com/journal/jems

|

335

Received: 3 May 2019

|

Revised: 13 October 2019

|

Accepted: 27 January 2020

DOI: 10.1111/jems.12342

ORIGINAL ARTICLE

Strategic shirking in competitive labor markets: A general

model of multi‐task promotion tournaments with

employer learning

Jed DeVaro

1

|Oliver Gürtler

2

1

Departments of Management and

Economics, California State University,

East Bay, Hayward, California

2

Department of Economics, University of

Cologne, Cologne, Germany

Correspondence

Oliver Gürtler, Department of Economics,

University of Cologne, Cologne, Germany.

Email: oliver.guertler@uni-koeln.de

Abstract

In a multitask, market‐based promotion tournament model, under different

environments concerning employer learning about worker ability, it is shown

that: (a) asymmetric learning in multitask jobs is a necessary condition for

“strategic shirking”(i.e., underperforming on certain tasks to increase the

promotion probability); (b) when learning becomes increasingly symmetric on

one task, the effort allocated to that task could increase or decrease, but effort

on the other task increases; (c) strategic shirking does not occur in equilibrium

in single‐task models; and (d) promotions signal worker ability even when

there is symmetric learning on one task, if learning is asymmetric on another.

1|INTRODUCTION

Ambitious, career‐conscious workers are ever alert to ways they can outshine their peers in promotion contests. Talent

and hard work are well‐known ingredients for success, but their roles in determining promotions complexify when jobs

are comprised of multiple tasks, as is usually the case in organizations. For example, police officers engage in primary

tasks (e.g., making traffic stops, and pursuing and apprehending suspects) and, to a lesser extent, in managerial and

leadership tasks, and public speaking. For police chiefs, however, the relative emphases placed on those activities are

reversed. Similarly, school teachers are primarily engaged with educating students, whereas managerial and leadership

activities (beyond the classroom setting) are not central to their job. For the school principals who supervise them,

however, the relative emphases on those tasks are reversed. In both settings, the police officers and teachers who do

stellar work on their primary tasks might actually be denied promotions because they are deemed indispensable in their

current jobs, relative to their coworkers. A better strategy for achieving promotion to police chief, or to principal, might

be to purposely underperform (or “strategically shirk”) on the primary tasks while overperforming on the leadership

tasks that are emphasized in the managerial job.

1

Strategic shirking is a fundamental incentive problem arising in multitask promotion tournaments, but it has only

been analyzed in “classic”settings in which the employer precommits to the wages paid to promoted and nonpromoted

workers to elicit desired worker behavior (e.g., effort choices) without direct influence from an outside market of

competing firms. The goal of this study is to analyze the problem in “market‐based”settings in which wages are

determined by the bids of competing firms, as in Waldman (1984). This contribution in a multitask setting parallels

recent developments in the (single‐task) tournament literature. The bulk of that literature, starting with Lazear and

Rosen (1981), concerns “classic tournaments”in which the employer can precommit to the prize structure. Recently, a

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

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.

© 2020 The Authors. Journal of Economics & Management Strategy published by Wiley Periodicals, Inc.

literature on (single‐task) “market‐based tournaments”has emerged, based on the realistic idea that the bids of

competing employers make precommitment challenging, if not impossible.

2

Moving from a classic to a market‐based wage‐setting mechanism is considerably more challenging in the multitask

environment than in the single‐task environment that the previous literature has addressed. One reason is that the

nature of the process by which competing employers learn about workers’talents becomes more complicated when

there are multiple tasks. Learning is crucial in a market‐based setting, because external compensation offers hinge on

what information potential employers learn about a worker’s abilities, which will often be less than the information

available to the worker’s current employer. When jobs are comprised of multiple tasks, however, many possible

learning environments arise. Competing employers might learn about primary‐task performance and leadership‐task

performance at the same rate as the worker’s current employer, as in the case of pure symmetric learning. Alternatively,

the current employer might learn performance on both tasks, whereas competing employers only learn that worker’s

job assignment, as in the case of pure asymmetric learning. Learning might also be symmetric on one task and

asymmetric on the other. Many intermediate cases are possible, involving mixtures of symmetric and asymmetric

learning, in which either or both tasks are observed by competing employers with certain probabilities.

The forthcoming analysis accommodates a range of possible learning environments. The theoretical model is quite

general, integrating the core features from the two pillars of the theoretical promotions literature (i.e., tournament

models and learning‐based job‐assignment models), both of which are heavily focused on single‐task jobs. Most models

in the promotions literature can be classified based on the following typology of features:

1. Fixed versus flexible managerial job slots.

3

2. Single‐task versus multitask jobs.

3. Endogenous worker choices (e.g., effort) versus not.

4. Heterogeneous versus homogeneous worker ability.

5. Symmetric versus asymmetric employer learning about worker ability.

4

6. Single firm versus multiple firms.

7. Classic versus market‐based wage setting.

The present model fully incorporates all of these dichotomies, assuming fixed managerial job slots in the main

analysis, with flexible slots covered in an extension. The model nests learning that is purely asymmetric, purely

symmetric, or a blend of both types, and it emphasizes multitask jobs while considering single‐task jobs in an extension.

Both worker ability and endogenous effort choices are incorporated and are task‐specific. Although the main focus is on

market‐based wage setting with multiple competing firms, another extension addresses classic wage setting in a single‐

firm context.

The model is the first in the literature to analyze task‐specific efforts and abilities in multitask job hierarchies in

competitive labor markets, under a range of assumptions about employer learning. Such a general model contributes in

three ways. First, it allows results on strategic shirking, which have recently been developed in the context of single‐

firm models in DeVaro and Gürtler (2016a,2016b), to be extended to more realistic and complex environments in

which multiple employers compete in the labor market under various assumptions about employer learning. Second, it

generates new insights, beyond simply the sum of those implied by both pillars of the theoretical promotions literature

considered separately. Third, it provides an organizing framework for understanding the literature and how various

theoretical results relate to particular modeling assumptions.

Concerning the first of these contributions, a key difference between a single‐firm environment and a multifirm

environment is information revelation, that is, how much information about worker ability do competing firms possess

(versus incumbent firms) and what assumptions govern how further information is revealed to competing firms over

time? Such considerations involving information allocation and revelation are absent in the single‐firm case and drive

the competitive wage bidding in the multiple‐firm case. When there are multiple firms, wage determination involves a

process of competitive bidding, which is more realistic than the single‐firm case in which the incumbent employer

chooses the wage.

Consider a labor market of competing employers, each of which has a two‐level hierarchy involving subordinate

jobs and managerial jobs. Suppose that two tasks (e.g., primary and leadership) are used in each of the two jobs, though

with primary tasks emphasized relatively more in the subordinate job’s production function. Those subordinates who

perform sufficiently well are promoted to managers, where they enjoy higher pay, whereas the others remain as

subordinates. Two questions arise concerning equilibrium behavior. First, do subordinates strategically shirk on

336

|

DEVARO AND GÜRTLER

primary tasks? Second, do they exert more effort, or less, when the learning environment becomes increasingly

symmetric and less asymmetric?

The answer to the first question is yes. Subordinates strategically shirk on primary tasks while hustling to perform

exceptionally well on leadership tasks. They engage in such distortion of effort across the two tasks because of their

employer’s inclination to promote the worker who has demonstrated the potential to be more productive as a manager

than as a subordinate. Interestingly, the nature of employer learning is shown to matter for the preceding result. In

particular, if learning is purely symmetric (i.e., symmetric on both tasks), then strategic shirking does not occur on

either task. Thus, asymmetric learning is a necessary, but not sufficient, condition for strategic shirking in multitask,

market‐based promotion models.

Concerning the second question, the degree to which effort allocation is inefficient depends on the employer

learning environment. The reason is that workers wish to impress potential employers if those employers are able to

observe workers’performance, but also that the learning environment has implications for wages, and wage differences

between job levels create incentives to strategically shirk. When the learning about ability on one task becomes

increasingly symmetric, the effect on that task’s effort is ambiguous, but effort on the other task increases. It is possible

that efficiency in effort allocation may decline as learning about ability becomes increasingly symmetric.

A policy implication is that employer practices designed to increase effort and that are frequently coupled with

promotion‐based incentives (e.g., incentive compensation) might exacerbate an inefficiency in effort allocation, for

example, by inducing workers to overwork on certain tasks. In the examples from the beginning of this section,

incentive compensation could be detrimental on leadership tasks.

5

Moreover, this result overturns the natural intuition

that labor market policies, technological change, or other institutional forces that move the labor market further from

asymmetric learning and closer to symmetric learning must increase welfare.

Whereas the preceding results concern inefficiency in effort allocation, the job‐assignment literature focuses instead

on inefficiency in job assignments. A celebrated result from that literature is that job assignments (in a single‐task

setting) are inefficient when employer learning about worker ability is asymmetric, and efficient when employer

learning is symmetric.

6

The main model of the present analysis assumes a fixed managerial job slot that must be staffed;

under those assumptions, job assignments are fully efficient. In an extension in which managerial job slots are flexible,

however, it is shown that the result concerning inefficient job assignments from the prior literature extends to a more

general, multitask setting with multidimensional abilities and efforts, as long as there is asymmetric learning on at least

one task.

Following the presentation of the model and its extensions, some empirical implications are highlighted, the most

interesting of which concerns overlapping wage distributions across adjacent job levels, which is an important phe-

nomenon that prior, influential theoretical models have been unable to fully explain.

2|BACKGROUND AND RELATED LITERATURE

Four prior studies provide empirical support for fixed job slots, as are assumed in the managerial job in the main

analysis. In a pair of related studies using samples of employers from four U.S. metropolitan areas, DeVaro

(2006a,2006b) provide evidence that relative performance determines promotions, which is a direct implication of fixed

managerial slots in a job hierarchy. In other words, when a worker’s competitors perform better, that worker’s

promotion chances suffer. In both of those studies, a worker’s promotion probability is found to be increasing in that

worker’s employer‐reported subjective performance rating and decreasing in the employer‐reported subjective per-

formance rating of the “typical worker”in that worker’s same job (which is taken as a proxy for the performance of

those against whom that worker competes for promotion) when both performance variables are simultaneously in-

cluded in the statistical model. DeVaro and Kauhanen (2016) find the same result using matched, employer‐employee

panel data from Finland. In contrast to the earlier two papers, DeVaro and Kauhanen (2016) do not rely on employer‐

reported subjective performance ratings but rather derive inferred performance measures (both for the worker in

question and for all other workers against whom that worker competes for promotion) from data on performance‐based

pay. Focusing on hiring rather than promotions, Lazear, Shaw, and Stanton (2018) extensively discuss fixed slots and

their empirical implications. They also use four data sets to provide an array of empirical results consistent with fixed

slots. First among these, and echoing the prediction from the three preceding studies of promotions, the authors find in

the oDesk data that the hiring probability depends not only on an applicant’s skills but also on the skills of the

competition.

7

DEVARO AND GÜRTLER

|

337