Misaccounting for endogeneity: The peril of relying on the Heckman two‐step method without a valid instrument

• Published date: 01 March 2019
• Author: Jordan Siegel,Sarah E. Wolfolds
• DOI: http://doi.org/10.1002/smj.2995
• Date: 01 March 2019

RESEARCH ARTICLE

Misaccounting for endogeneity: The peril of relying

on the Heckman two-step method without a valid

instrument

Sarah E. Wolfolds

1

| Jordan Siegel

2

1

Charles H. Dyson School of Applied

Economics & Management, Cornell SC Johnson

College of Business, Ithaca, New York

2

Ross School of Business, University of Michigan,

Ann Arbor, Michigan

Correspondence

Sarah E. Wolfolds, Charles H. Dyson School of

Applied Economics & Management, Cornell SC

Johnson College of Business, 375E Warren Hall,

Ithaca, NY 14853.

Email: sew276@cornell.edu

Funding information

Cornell University Dyson School of Applied

Economics and Management, University of

Michigan Ross School of Business

Research Summary:Strategy research addresses endo-

geneity by incorporating econometric techniques, includ-

ing Heckman's two-step method. The economics literature

theorizes regarding optimal usage of Heckman's method,

emphasizing the valid exclusion condition necessary in

the first stage. However, our meta-analysis reveals that

only 54 of 165 relevant papers published in the top strat-

egy and organizational theory journals during 1995–2016

claim a valid exclusion restriction. Without this condition

being met, our simulation shows that results using the

Heckman method are often less reliable than OLS results.

Even where Heckman is not possible, we recommend that

other rigorous identification approaches be used. We illus-

trate our recommendation to use a triangulation of identifi-

cation approaches by revisiting the classic global strategy

question of the performance implications of cross-border

market entry through greenfield or acquisition.

Managerial Summary:Managers make strategic deci-

sions by choosing the best option given the particular cir-

cumstances of their firm. However, researchers had

previously not taken into consideration these circum-

stances when evaluating the outcome of that choice. The

Heckman method importantly addresses this situation, but

requires that the researcher have some variable that effects

the best option for the firm, but not the outcome. We show

that researchers frequently do not utilize such a variable,

and demonstrate that the Heckman method can exacerbate

estimation issues in this case. We then provide an

approach that researchers can use to address the challenge

of determining the outcome of a strategic decision, and

illustrate it with an empirical examination of the perfor-

mance implications of cross-border market entry through

greenfield or acquisition.

Received: 9 November 2017 Revised: 24 September 2018 Accepted: 15 November 2018

DOI: 10.1002/smj.2995

432 © 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/smj Strat Mgmt J. 2019;40:432–462.

KEYWORDS

entry mode, foreign investment, Heckman correction,

selection bias, two-step estimator

1|INTRODUCTION

Strategy research has recently encouraged the use of advanced econometric techniques such as the

Heckman two-step method. Though econometric methods can be vital to draw conclusions from non-

random data, understanding the theoretical underpinnings, and especially the assumptions, that go

into them is crucial, particularly given that newer programming languages contain precoded packages

and routines that require little knowledge of the model's mechanics. Recent papers in Strategic Orga-

nization have even provided STATA code to perform Heckman's two-step method. Bascle (2008)

and Hamilton and Nickerson (2003) both provide concise presentations of the Heckman two-step

method as it applies to strategic management research, whereas Certo, Busenbark, Woo, and Sema-

deni (2016) provide guidelines for determining when the Heckman method is not necessary.

A key tenet of the strategy field is that firms are heterogeneous and that managers will choose a

strategy based on this heterogeneity. For example, managerial decisions are often endogenous to firm

performance (Bascle, 2008) and, similarly, the resource-based view asserts that, to obtain competitive

advantage, firms ought to diversify and develop products based on their unique resources. In other

words, “endogeneity bias is a concern only when firms have some unobservable (to the researcher)

advantage or disadvantage that influences the strategy they choose”(Hamilton & Nickerson, 2003,

p. 65). Because much of strategy research stipulates that a strategy or resource needs to be both inimi-

table and uniquely based on firm characteristics to sustain a competitive advantage, it seems reason-

able to conclude that endogeneity is a widespread concern in this area.

Thus, the Heckman method can be especially useful in strategic management research, where endo-

geneity is a frequent problem to solve. However,the Heckman method should not be used without care-

ful consideration of its appropriateness and of the ability to implement it accurately. Our paper analyzes

a way in which the Heckman method is often implemented that lowers its reliability: without a valid

exclusion restriction in the first stage, described in Section 2. In Section 3, we use simulation analysis

to illustrate how the Heckman method compares to OLS when used with and without a valid exclusion

restriction, finding that OLS performs better than the Heckman method without a valid exclusion

restriction when the assumption that the errors follow a bivariate normal distribution fails to hold. This

is a relevant concern because the assumption that large sample size will assuage concerns of non-

normality fails in a majorityof cases due to lumpiness and asymmetry in data (Micceri, 1989). We then

illustrate the prevalence of this issue via a meta-analysis of the strategy and organizational theory litera-

tures in Section 4, showing that the Heckman method continues to be used without a valid exclusion

restriction. Finally, in Section 5, we revisit a classic global strategy question, namely whether there is

an average performance advantage between greenfield-based foreign entry and acquisition-based for-

eign entry, to demonstrate the triangulation of approaches we suggest to future researchers. Shaver

(1998) utilized the same question to introduce the Heckman method to the strategy literature; we

approach the question with different data to provide practical evidence on this important methodological

point. Section 6 concludes and provides suggestions for future researchers.

WOLFOLDS AND SIEGEL 433