Addressing the endogeneity dilemma in operations management research: Theoretical, empirical, and pragmatic considerations

• Author: Mikko Ketokivi,Cameron N. McIntosh
• Date: 01 May 2017
• Published date: 01 May 2017
• DOI: http://doi.org/10.1016/j.jom.2017.05.001

Addressing the endogeneity dilemma in operations management

research: Theoretical, empirical, and pragmatic considerations

Mikko Ketokivi

a

,

*

, Cameron N. McIntosh

b

a

IE Business School, IE University, Spain

b

Government of Canada, Public Safety, Canada

article info

Article history:

Received 25 September 2016

Accepted 12 May 2017

Available online 27 May 2017

Handling Editor: Dr. V. Daniel R. Guide

abstract

In this paper, we examine the problem of endogeneity in the context of operations management

research. Whereas the extant literature has focused primarily on the statistical aspect of the problem, a

comprehensive treatment requires an examination of theoretical and pragmatic considerations as

complements. The prevailing problem with the focus on statistical techniques is that the standards tend

to be derived from idealizations: the correlation between a regressor and a disturbance term must be

exactly zero, or the analysis will be invalid. In actual empirical research settings, such a knife-edge

assumption can never be satisfied, indeed it cannot even be directly tested. Idealizations are useful in

helping us understand what it would take to eliminate endogeneity, but when applied directly and

unconditionally, they lead to unreasonable standards that may unnecessar ily stifle substantive inquiry.

We believe that it is far more productive and meaningful to ask: “What can we realistically expect

empirical scientists to be able to achieve?”To this end, we cover and revisit someof the generaltechnical

material on endogeneity, paying special attention to the idiosyncrasies of operations ma nagement

research and what could constitute reasonable criteria for addressing endogeneity in empirical opera-

tions management studies.

©2017 Elsevier B.V. All rights reserved.

1. Introduction

Operations management (OM) researchers apply a variety of

methods and research designs, but statistical modelingdanalysis of

variance, regression analysis, factor analysis, and structural equa-

tion modelingdremains by far the most commonly used tool we

use to make inferences and to draw conclusions from empirical

data. Despite decades of empirical OM research and much progress,

there are a number of key assumptions underlying our statistical

reasoning techniques that have not been sufficiently well expli-

cated. These assumptions, when violated, can have fundamental

implications for the credibility of our inferences and our theoretical

interpretations. In this paper, we aim to explore some of the most

insidious threats to trustworthy statistical inference, as well as the

various ways in which researchers can tackle these threats.

To set the stage for our inquiry, consider the link between plant

productivity and business unit profitability. Even a casual theoret-

ical reflection suggests that as plant productivity rises, profitability

rises as well; a simple regression analysis would confirm the rela-

tionship to be positive. But to what extent do increases in pro-

ductivity actually drive increases in profitability? The answer to this

question requires that we get not just the sign but also the

magnitude of this effect right. In this paper, we discuss what is

perhaps the most significant threat to getting the magnitude right.

In the econometrics literature, this is dubbed the problem of endo-

geneity. Sometimes endogeneity causes so much bias that we may

not even get the sign of the coefficient right.

In order to bring further conceptual and statistical clarity to the

endogeneity problem, let us put the productivity-profitability

relationship into a model by supposing a fairly typical OM

research model, depicted in Fig. 1:

1)

x

1

-x

3

are aspects of a factory's production system, such as

process choice or the degree of implementation of certain

manufacturing principles and practices;

2)

y

1

is a measure of the factory's operational performance, say,

total factor productivity; and

3)

y

2

is a measure of business success, say, business unit

profitability.

*Corresponding author.

E-mail addresses: Mikko.Ketokivi@ie.edu (M. Ketokivi), Cameron.McIntosh@

aadnc-aandc.gc.ca (C.N. McIntosh).

Contents lists available at ScienceDirect

Journal of Operations Management

journal homepage: www.elsevier.com/locate/jom

http://dx.doi.org/10.1016/j.jom.2017.05.001

0272-6963/©2017Elsevier B.V. All rights reserved.

Journal of Operations Management 52 (2017) 1e14