Int. J.

Production Economics 138 (2012) 242253

Contents lists available at SciVerse ScienceDirect

Int. J. Production Economics

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

The effect of lean production on nancial performance: The mediating role

of inventory leanness
Christian Hofer a,n, Cuneyt Eroglu b, Adriana Rossiter Hofer a
a
Department of Supply Chain Management, Sam M. Walton College of Business, University of Arkansas, 475 Business Building, Fayetteville, AR 72701, USA
b
Information, Operations & Analysis Group, College of Business Administration, Northeastern University, 214 Hayden Hall, 360 Huntington Avenue, Boston, MA 01115, USA

a r t i c l e i n f o a b s t r a c t

Article history: The purpose of this paper is to empirically investigate the relationship between lean production
Received 28 September 2011 implementation and nancial performance. Particular emphasis is placed on the mediating role of
Accepted 23 March 2012 inventory leanness in deriving the nancial performance benets commonly associated with lean
Available online 30 March 2012
production. Moreover, the interaction among different lean practice bundles in affecting nancial and
Keywords: inventory performance is assessed. Based on an analysis of a combination of survey and secondary data,
Lean production the effect of lean production on nancial performance is found to be partially mediated by inventory
Inventory management leanness. In addition, there is strong evidence that the concurrent implementation of internally-focused
Financial performance and externally-focused lean practices yields greater performance benets than selective lean produc-
tion implementation. Thus, this study contributes to the theory of lean production by providing insights
into the mediated and moderated effects of lean production on inventory leanness and nancial
performance.
& 2012 Elsevier B.V. All rights reserved.

1. Introduction
Lean production is often regarded as the gold standard of modern
operations and supply chain management (e.g., Guinipero et al.,
2005; Goldsby et al., 2006). Numerous studies have investigated the
relationship between lean production and nancial performance
(e.g., Fullerton et al., 2003; Jayaram et al., 2008). Yet, the exact
mechanism(s) through which lean production affects nancial
performance remain underresearched. Conventional wisdom holds
that, as a manufacturing strategy, lean production strives to mini-
mize waste and thereby increase efciency (Womack et al., 1990),
and by extension, nancial performance.
Given the multiplicity of lean production practices such as
kanban, JIT, and TQM, for example, it is apparent that the relationship
between lean production and nancial performance may be complex
and multi-faceted. Indeed, one factor that is often implicitly con-
sidered as a mediator of this relationship is inventory efciency. For
example, several studies have examined the effects of lean produc-
tion implementation on inventories (e.g., Huson and Nanda, 1995;
Balakrishnan et al., 1996). Likewise, analytical research has examined
the linkage between production and inventory (e.g., Miyazaki, 1996;
Dobos, 2007). In a separate literature stream, prior research has
investigated the performance implications of efcient inventory
n
Corresponding author. Tel.: 479 575 6154.
E-mail address: chofer@walton.uark.edu (C. Hofer).
management (e.g., Capkun et al., 2009; Eroglu and Hofer, 2011).
Moreover, Fullerton and Wempe (2009) contend that the effects of
lean production implementation on nancial performance are
mediated by various operational performance measures, such as
delivery performance, manufacturing cycle times, and labor produc-
tivity. However, these authors do not consider inventories as a
mediating factor. Yet, inventory costs are of great signicance in
the context of logistics and supply chain management (Stock and
Broadus, 2006).
Thus, the purpose of this study is to add to our understanding
of lean production by examining the relationship between lean
production and nancial performance, with an emphasis on the
mediating role of inventories. In addition, and consistent with the
notion that lean production is a system of lean practices (Womack
et al., 1990), interactions among various facets of lean production
and their effects on inventories and performance are investigated.
This research contributes to the existing literature in multiple
ways: First, it provides a richer, more nuanced conceptualization of
the relationships among lean production, inventory leanness, and
nancial performance. Specically, we draw on existing lean
production and inventory literature to develop a research model
that examines the mediating role of inventory in delivering the
commonly expected nancial performance benets of lean produc-
tion implementation. This model is tested using a data sample with
rm-level observations from a diverse set of US manufacturing
industries which is compiled from two distinct sources: survey
data and matched secondary nancial data. Beyond conventional

0925-5273/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ijpe.2012.03.025
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253
mediation analysis, we also test for potential reverse causality and,
thus, gain a better understanding of the interplay of lean produc-
tion, inventory leanness, and nancial performance.
Second, this study explores interactions among lean practices.
Existing literature suggests that when lean practices are imple-
mented concurrently, the total performance effect will exceed the
sum of performance effects of individual lean practices (Shah and
Ward, 2003). While some recent studies (e.g., Furlan et al., 2011a)
have empirically tested the complementarity (synergy) among
lean practice bundles, these analyses were restricted to specic
aspects of lean production and focused on plant-level perfor-
mance. In this study, we conceptualize lean production as two
lean practice bundles (internal and external) that collectively
encompass all lean practices, and we test the synergy between
these lean practice bundles at the rm-level instead of at the
plant level. In addition, this research addresses concerns of
potential common methods bias (Podsakoff et al., 2003) that
may arise when lean production and performance data are
provided by the same survey respondent by using secondary
inventory and nancial performance metrics along with primary
survey data on lean production implementation.
Third, the two main constructs of this study, i.e., lean production
and inventory leanness, are operationalized using measures pro-
posed in recent research. More specically, lean production is
assessed using a survey instrument developed by Shah and Ward
(2007) which consists of a set of 10 distinct lean practices: supplier
feedback, supplier JIT, supplier development, customer involvement,
pull manufacturing, continuous ow manufacturing, setup time
reduction, statistical process control, employee involvement, and
total productive maintenance. Inventory leanness, in turn, is mea-
sured using the Empirical Leanness Indicator (ELI) developed by
Eroglu and Hofer (2011). The ELI measures a rms inventory
leanness as the deviation of a rms inventory levels from size-
adjusted within-industry average inventory levels. As such, the study
extends operations management literature by providing independent
empirical evidence for the validity of these instruments.
The remainder of this paper is structured as follows: The
relevant literature is reviewed and hypotheses are proposed in
Section 2. Data and measurement issues are discussed in Section 3.
In Section 4, mediation and interaction hypotheses are tested and
empirical estimation results are presented. Section 5 presents a
discussion of the ndings, research and managerial implications,
limitations, and future research opportunities.
2. Literature review and hypothesis development
Lean production is a strategy or philosophy that promotes the
use of practices, such as kanban, total quality management (TQM)
and just-in-time (JIT), to minimize waste and enhance rm
performance (Womack et al., 1990). Thus, the implementation
of lean production practices is expected to result in improved
operational outcomes, such as lower inventories, higher quality,
and shorter throughput times, which, in turn, should improve
nancial performance. This description of lean production clearly
indicates a number of mediating factors between lean production
and nancial performance. This notion is consistent with the
new inventory paradigm (Chikan, 2011, 2009) which empha-
sizes the connectedness to other processes and functions within
rms and to rm protability.
In this study, we focus on inventory leanness as the mediator
of interest and suggest that inventory leanness mediates the
effect of lean production implementation on nancial perfor-
mance. Inventory leanness is dened by comparing a rms
inventory levels to the size-adjusted average inventory level
within the rms industry (Eroglu and Hofer, 2011). It is expected
243
that lean production implementation not only carries direct
nancial benets, but also results in greater inventory leanness
which, in turn, contributes to improved nancial performance.
In accordance with the proposed research model, relevant
literature is grouped in three distinct streams as shown in Fig. 1
adapted from Fullerton and Wempe (2009). The rst stream
consists of studies exploring the relationship between lean produc-
tion and nancial performance. The second stream examines the
effect of lean production on inventory leanness and other opera-
tional outcomes as potential mediators. The third stream focuses
on the analysis of the relationship between inventory leanness and
nancial performance. All three streams are reviewed below,
followed by the presentation of research hypotheses on the
mediating role of inventory leanness in the lean production-
nancial performance relationship and the interaction effects
among lean production practice bundles.
Stream 1
Lean Financial
Production Performance
Stream 2 Stream 3
Inventory
Leanness
Fig. 1. Research streams on lean production and rm performance.
2.1. Stream 1: Relationship between lean production and nancial
performance
The rst stream of research explores the direct effects of lean
production practices on nancial performance (Table 1). Most of
these studies employ a survey methodology to assess the degree
of implementation of lean production practices and to measure
nancial performance. The measures of lean production are
typically narrowly focused on JIT (e.g., Inman and Mehra, 1993;
Fullerton and McWatters, 2001) which is part of but not synon-
ymous with lean production. Other studies identify companies
that have adopted JIT practices via a search of news articles and
company reports (Biggart, 1997; Kinney and Wempe, 2002). Firm
nancial performance, in turn, is estimated using metrics such as
ROS, ROA, and ROI in most studies.
The studies ndings are largely consistent: Evidence of positive
effects of lean production implementation on at least some nancial
performance indicators is presented by Inman and Mehra (1993),
Callen et al. (2000), Fullerton and McWatters (2001), Germain et al.
(1996), Kinney and Wempe (2002), Fullerton et al. (2003), Fullerton
and Wempe (2009), and Yang et al. (2011). Only Biggart (1997) and
Jayaram et al. (2008) nd no statistically signicant relationships
between lean production practices and rm protability.
2.2. Stream 2: Relationship between lean production and inventory
leanness
The second stream of research explores the effects of lean
production implementation on inventory leanness and other
operational performance measures (Table 2). Most of the studies
in this literature stream are based on surveys of manufacturing
executives (e.g., White, 1993; Norris et al., 1994; Droge and
Germain, 1998; Shah and Ward, 2003). These studies typically
employ multi-item scales to measure the degree of implementa-
tion of lean production. Inventory and operational performance,
244 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253

Table 1
Studies on the lean productionnancial performance relationship (Stream 1).

Author (s) Sample Dependent Independent Empirical Findings

variable(s) variable(s) methodology

Inman and US manufacturing rms ROI, total cost, JIT adoption Regression Firm performance improves as a result of JIT adoption.
Mehra (1993) adopting JIT (N 114) service
Biggart (1997) US manufacturing rms ROA JIT adoption Regression No evidence of a signicant effect of lean production
adopting JIT (N 106) adoption on ROA is found.
Claycomb et al. US manufacturing ROS, ROI, prot, JIT adoption Regression JIT use with customers results in better nancial
(1999) managers (N 200) prot growth performance.
Claycomb et al. US manufacturing ROS, ROI, prot JIT adoption Regression The greater the share of JIT transactions, the greater ROI,
(1999) managers (N 200) ROS, and rm protability.
Callen et al. Canadian manufacturing Protability, total JIT adoption Regression JIT adoption results in lower costs and higher prots.
(2000) plants (N 100) costs
Fullerton and US manufacturing rms Protability JIT adoption ANOVA Greater JIT implementation results in greater protability
McWatters adopting JIT (N 95) improvement improvement.
(2001)
Germain et al. US manufacturing ROS, ROI, prot JIT adoption Regression JIT results in greater nancial performance relative to
(1996) managers (N 200) industry peers.
Kinney and US manufacturing rms Protability, ROA JIT adoption Regression Protability and return on assets improve after JIT adoption.
Wempe (2002) adopting JIT (N 201  2)
Fullerton et al.
Manufacturing rms Protability, cash Lean production Regression Three lean production practice bundles are associated with
(2003) (N 253) ow margin, ROA implementation greater rm performance.
Matsui (2007) Japanese manufacturing Manufacturing cost JIT adoption Canonical JIT production systems contribute to competitive
plants (N 46) correlations performance outcomes such as lower manufacturing costs.
Jayaram et al. Auto parts manufacturers Protability, ROA Lean production SEM Firm performance is not signicantly affected by lean
(2008) (N 57) implementation production.
Fullerton and Manufacturing executives ROS Lean production SEM Lean practices have a direct and mediated positive effect on
Wempe (2009) (N 121) implementation nancial performance.
Yang et al. IMSS survey data (N 309) ROS, ROA Lean SEM Lean manufacturing has a signicant positive impact on
(2011) manufacturing nancial performance.

Table 2
Studies on the lean productioninventory leanness relationship (Stream 2).

Author (s) Sample Dependent variable(s) Independent Empirical Findings

variable(s) methodology

White (1993) Manufacturing and Throughput time JIT adoption Percentage About 50% of the respondents report decreased throughput
service rms (N 1035) breakdown times after JIT implementation.
Norris et al. Plant managers (JIT Inventory performance, JIT adoption Percentage Respondents report lower inventories and greater inventory
(1994) users) (N 48) process/quality control breakdown visibility and accuracy following JIT implementation.
items
Huson and US manufacturing rms Inventory turnover JIT adoption Simultaneous Inventory turnover increases following lean production
Nanda (N 55) equations implementation.
(1995)
Balakrishnan US manufacturing rms Inventory turnover JIT adoption t-test Firms that adopt JIT achieve higher inventory turnover than
et al. (N 46  2) rms that do not use lean production.
(1996)
Droge and US manufacturing Inventory JIT adoption Correlation A negative correlation between JIT adoption and inventory
Germain managers (N 200) analysis levels is found.
(1998)
Irvine (2003) Aggregate US inventory Inventory-to-sales ratio JIT adoption Descriptive Descriptive evidence suggests that JIT adoption has resulted in
and sales data (34 years) analysis large inventory reductions in the durable goods manufacturing
sector.
Biggart and US manufacturing rms Inventory-to-sales ratio JIT adoption t-test A decrease in total and raw materials inventories is observed
Gargeya (N 74) after JIT implementation.
(2002)
Shah and Manufacturing plants Operational performance Lean Regression Lean production practice bundles have a positive effect on
Ward (N 1575) construct production operational performance.
(2003) implementation
Demeter and International Inventory turnover Lean Cluster Firms that implement lean practices have higher inventory
Matyusz Manufacturing Strategy production analysis, turnover.
(2011) Survey (N 711) implementation F-tests

in turn, is measured in terms of inventory levels (e.g., Droge and
Germain, 1998), cycle time (White, 1993) or as a multi-dimen-
sional construct including elements such as inventory visibility
and accuracy as well as quality and process control (Norris et al.,
1994; Shah and Ward, 2003). Even though there are some
differences in variable measurement and survey populations, all
studies present evidence that greater implementation of lean
production is associated with improvements in at least some
aspects of inventory and operational performance. These ndings
are further corroborated by a set of studies that identify the
adoption of JIT practices, a subset of lean production practices, via
a literature search and rely on secondary data only to estimate the
effect of JIT adoption on inventory performance (Huson and
Nanda, 1995; Balakrishnan et al., 1996; Biggart and Gargeya,
2002). Specically, these studies nd that JIT adoption leads
to increased inventory turnover (Huson and Nanda, 1995;
Balakrishnan et al., 1996) and lower raw materials inventories
(Biggart and Gargeya, 2002). It is noteworthy, however, that only
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253 245

one study has explored the effect of lean production practices
other than JIT adoption on inventory leanness (Demeter and
Matyusz, 2011). Based on a series of univariate tests, these
authors conclude that raw materials, work-in-process and
nished goods materials inventories are lower when the company
has implemented lean practices. Consistent with these rm-level
studies, aggregate analyses of inventory and sales data for also
indicate that the implementation of JIT has resulted in lower
inventory holdings (Irvine, 2003; Obermaier, 2012).
2.3. Stream 3: Relationship between inventory leanness and
nancial performance
of assemble-to-order production systems contributes to cost
savings through reduced work-in-process and raw materials
inventory requirements. Fullerton and Wempe (2009), in turn,
conducted a survey study and found evidence that non-nancial
manufacturing performance measures, such as on-time deliveries
and labor productivity, mediate the relationship between lean
production implementation and ROS. While neither study empiri-
cally assessed the mediating role of inventories, both studies
support the contention that operational outcomes, such as inven-
tory leanness, act as mediating variables in the relationship
between lean production implementation and nancial perfor-
mance. Thus, it is hypothesized that:

The third stream of research analyzes the link between
inventory leanness and nancial performance (Table 3). There is,
of course, a large body of conceptual and analytical literature that
discusses the prot implications of inventories (e.g., Relph and
Barrar, 2003; Chikan, 2011). Empirical studies in this stream of
research use large cross-sectional time series data sets compiled
from secondary databases to explore the link between inventory
metrics and nancial outcomes. Chen et al. (2005, 2007) nd that
rms with inventory levels below industry average tend to have
greater stock returns. Similarly, Eroglu and Hofer (2011) nd that
rms whose inventory levels are (slightly) below size-adjusted
industry averages tend to exhibit greater nancial performance.
Swamidass (2007), Koumanakos (2008) and Capkun et al. (2009)
also present evidence of a negative relationship between inven-
tory levels and protability. Cannon (2008), however, nds no
signicant relationship between inventory turnover and nancial
performance.
2.4. The mediating effect of inventory leanness
As outlined above, the existing literature documents the effect
of select aspects of lean production on nancial performance as
well as inventories and other measures of operational perfor-
mance. The underlying logic is that operational benets typically
associated with lean production, such as lower inventories, higher
quality, and less waste ultimately lead to improved nancial
performance. Mistry (2005) and Fullerton and Wempe (2009)
explicitly study the interplay between lean production, opera-
tional outcomes, and nancial performance. Based on a case study
of an electronics manufacturer that had previously implemented
select JIT processes, Mistry (2005) identied several processes
through which JIT practices impacted various operational
outcomes and, ultimately, increased protability. In addition,
Mistry (2005) found anecdotal evidence that the implementation
H1. The effect of lean production on nancial performance is
mediated by inventory leanness.
2.5. Interaction among lean practice bundles
Although lean production has traditionally been conceptua-
lized as a collection of lean practices (e.g., Shah and Ward, 2007),
it is implied that these distinct practices should work together as
a system (Womack et al., 1990). Several years after the publica-
tion of their seminal book on lean production, Womack and Jones
(1996, p. 140) observed many unlinked islands of lean operating
techniques and commented that [a]lthough many managers
had grasped the power of individual lean techniques quality
function deployment for product development, simple pull sys-
tems to replace complex computer systems for scheduling, and
the creation of work cells for operations ranging from credit
checking and order entry in the ofce to parts fabrication in the
plant they had stumbled when it came to putting them all
together into a coherent business system. That is, they could hit
individual notes (and loved how they sounded) but still could not
play a tune. Several conceptual studies have endorsed the view
that lean production is a system of inter-related practices (Huang,
1991; Roth and Miller, 1992; Imai, 1998).
A number of recent studies have empirically tested the
synergy among different lean practice bundles. Konecny and
Thun (2011) nd that while the adoption of TQM and TPM
bundles individually improves plant performance, their conjoint
implementation does not provide additional performance bene-
ts. Furlan et al. (2011b) detect synergy between JIT and TQM
bundles for rms that also implement human resources related
lean practices. In another study, Furlan et al. (2011a) show that
simultaneously implementing upstream and downstream JIT
practices improves plant performance to a greater degree than
implementing these practices separately. Although these studies

Table 3
Studies on the inventory leannessnancial performance relationship (Stream 3).

Author (s) Sample Dependent variable(s) Independent Empirical Findings

variable(s) methodology

Chen et al. US manufacturers Stock returns, Tobins q, Abnormal Linear mixed Firms with below average inventories have better stock
(2005) 19812000 (N 7,433) Market-to-book ratio inventory, time models returns than rms with more or less inventory.
Chen et al. Retailers, wholesalers Stock returns Abnormal Linear mixed High inventory levels are associated with poor long-term
(2007) 19812004 (N 1,662) inventory, time models stock returns.
Swamidass US manufacturers Inventory-to-sales ratio Z-score (rm Regression Top performers have decreasing inventory-to-sales ratio
(2007) 19811998 (N 14,400) performance), time over time, while low performers have increasing ratios.
Cannon US manufacturers Market value added, Inventory turnover, Hierarchical Inventory turnover has little or no effect on nancial
(2008) 19912000 (N 2,440) ROA, ROI, Tobins q capital intensity linear models performance.
Koumanakos Greek manufacturers Gross margin, net Inventory days Regression In most industries, greater inventory levels are associated
(2008) 20002002 (N 1,358) operating margin with lower net operating margins.
Capkun et al. US manufacturers Gross prot, EBIT Inventory scaled by Regression Greater inventory performance (raw materials, in particular)
(2009) 19802005 (N 52,254) sales positively affects rm performance.
Eroglu and US manufacturers ROS, ROA Empirical leanness Regression The relationship between inv. leanness and nancial
Hofer 20032008 (N 7,804) indicator (ELI) performance is concave in most industries.
(2011)
246 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253
do not directly analyze nancial performance and inventories,
they suggest that synergies exist among lean practice bundles.
Hence, it is hypothesized that:
H2. Lean practices interact positively to affect nancial performance.
H3. Lean practices interact positively to affect inventory leanness.
3. Data collection and measurement
To test the hypotheses set forth in this study, a data set
composed of primary survey data and secondary nancial data
was utilized. Primary survey data were collected to measure the
degree of a rms implementation of lean production, whereas
secondary nancial data were obtained to measure a rms
inventory leanness and nancial performance. The combination
of both primary and secondary data sources is intended to
address common methods bias which is associated with most
survey research (Podsakoff et al., 2003).
3.1. Survey data
Firm-level lean production was measured by administering a
survey instrument developed by Shah and Ward (2007). This
formative scale comprises 41 questions which capture 10 lean
production practices: supplier feedback, supplier JIT, supplier
development, customer involvement, pull manufacturing, contin-
uous ow manufacturing, setup time reduction, statistical process
control, employee involvement and total productive maintenance.
All questions were answered on a ve-point Likert scale ranging
from (1) no implementation to (5) complete implementation.
The survey items are listed in Table 11 in Appendix A.
The survey instrument was administered by the Association for
Operations Management (APICS) whose professional membership
base was deemed a particularly suitable sampling frame because
the majority of its members are operations, production, supply
chain, logistics, and purchasing managers and executives. Email
solicitations were sent by APICS to a mailing list comprising a
random sample of 4288 APICS members. Excluding 38% of APICS
members that were afliated with non-manufacturing organiza-
tions, the sampling frame consisted of about 2662 APICS members.
After an initial invitation email and a subsequent reminder email,
788 email recipients had opened the email and 325 individuals
clicked on the survey link contained in the email message. A total
of 229 responses were obtained, corresponding to a lower-bound
response rate of 8.6% (229/2662) relative to the number of APICS
members in the manufacturing sector. However, some of the 2662
email recipients in the manufacturing sector may not be qualied
to take the survey due to limited exposure to lean production
related activities within their respective organizations. Assuming
that all ineligible individuals excluded themselves and only eligible
participants opened the email message, we obtain an upper-bound
response rate of 29.1% (229/788). While the true response rate may
be difcult to pinpoint accurately, in all likelihood, it lies some-
where between 8.6% and 29.1%. Our survey effort, thus, compares
Table 4
Sample demographics.
Title n % Department
CEO/COO/CFO 2 1 Manufacturing/operations
VP/EVP/SVP 9 4 Logistics/SCM/procurement
Director 34 15 Other
Manager 89 39
Othera 95 41
a
Other titles include, most notably, supervisor, (production) planner, scheduler, analyst, and (senior) buyer.
well both in terms of response rate as well as in terms of absolute
sample size to other large scale survey studies in operations
management (e.g., Braunscheidel and Suresh, 2009; Hult et al.,
2007; Bardhan et al., 2007).
It is conceivable that email recipients declined to take the
survey due to limited or negative experiences with lean produc-
tion. As such, the possibility of non-response bias needs to be
considered. To do so, we rst follow the procedure suggested by
Armstrong and Overton (1977) and Lambert and Harrington (1990)
and compare demographic characteristics between early and late
respondents. No signicant differences were found. Specically,
the survey data of early respondents (rst quartile of respondents,
nq1 57) and late respondents (fourth quartile of respondents,
nq4 57), were compared using Hotellings T-squared test. This
test yielded an F statistic of 1.22 (p0.24). Hence, the null
hypotheses of equal vectors of means across early and late
respondents could not be rejected. Second, we identied publicly
traded rms among the respondents and compared their distribu-
tion across NAICS codes with the distribution of all publicly traded
rms in the COMPUSTAT database. A test of independence failed to
detect any signicant differences between the publicly traded
rms in the survey sample and those in the population of
COMPUSTAT (asymptotic chi-square test p 0.4174; Fishers exact
test p0.2092). These two ndings ease potential concerns of
non-response bias.
Sample demographics are provided in Table 4. A majority of
respondents held managerial and supervisory positions in man-
ufacturing, operations, logistics, procurement and supply chain
management. Likewise, 59% of all respondents had been with
their respective rms for more than ve years, while another 33%
had been with their employer for between one and ve years.
Based on these demographics the respondents were deemed
qualied to complete this survey. A conrmatory factor analysis
of the survey items was conducted and acceptable model t
statistics were obtained for all 10 lean production constructs
(Table 11 in Appendix A).
3.2. Secondary data
In addition to survey data, rm-level nancial and inventory
data were obtained from Standard & Poors COMPUSTAT database
for 2009 (which is also the year in which the survey data were
collected). As with the survey data collection, the query was
limited to US domestic manufacturing rms that were active and
had positive sales and inventory gures. The resulting data set
included 1421 rms in 24 four-digit NAICS manufacturing
industries.
Inventory leanness was measured using the Empirical Lean-
ness Indicator (ELI) developed by Eroglu and Hofer (2011). This
measure overcomes some of the shortcomings of measures used
in previous studies. Inventory turnover and its variants, while
widely used in the literature, are imperfect measures of inventory
leanness for two reasons. Since such measures are scaled by rm
size, they yield articially inated estimates when used with rm
performance measures that are also scaled by rm size or some
n % Tenure n %
78 34 Less than 1 year 18 8
133 58 1 to 5 years 76 33
18 8 More than 5 years 135 59
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253 247

highly correlated variable (Wiseman, 2009). In addition, such In line with previous research by Shah and Ward (2003), lean
measures ignore economies of scale in inventory management practices are grouped into lean practice bundles using principal
(Evers, 1995). More detailed information on the ELI and its component analysis and VARIMAX rotation (Table 6). These
computation is provided in Appendix B. We note that inventory results provide a factor structure that is both easy to interpret
leanness is not reective of lean production implementation. and theoretically meaningful: All externally-oriented lean pro-
Rather, it is a theoretically and empirically distinct variable that duction practices (supplier feedback, supplier JIT, supplier devel-
is expected to be impacted by lean production implementation opment, customer involvement) load on factor 1 (termed
and, at the same time, impact nancial performance. external lean practices, ELP) whereas all internally-oriented
Firm nancial performance was measured with return on sales lean production practices (pull system, continuous ow, setup
(ROS), a metric commonly used in prior research (Kinney and time reduction, statistical process control, employee involvement,
Wempe, 2002; Cannon, 2008; Koumanakos, 2008). It is noted, total productive maintenance) load on factor 2 (termed internal
however, that the results are largely insensitive to small changes lean practices, ILP). ELP and ILP factor scores were calculated as
in the measurement of nancial performance. Specically, the the average of the factors respective constituent items.
empirical results remain essentially the same when return on In order to test for discriminant validity between the con-
assets (ROA) is used as a nancial performance metric. These structs of internal and external lean practices, we followed the
results are not reported in this paper due to space constraints. methodology suggested by Shook et al. (2004). The shared
In line with prior research, sales and sales growth data are variance between factors should be lower than the average
included as control variables in the empirical model. Total sales variance extracted for each factor. A two-factor model was tested
are a measure of rm size, which has been shown to be associated assigning the items to their respective factors and allowing both
with larger prots (e.g., Barber and Lyon, 1996). Likewise, prior factors (internal and external lean practices) to covary. For both
research has established the higher sales growth is correlated factors, the variance extracted was higher than the covariance
with greater protability (e.g., Faireld and Yohn, 2001). shared by the factors, indicating discriminant validity. In addition,
the two factor model presented an excellent t (CFI.981.
SRMR0.040, RMSEA0.043), which according to Kline (2005) is
also a precise test of discriminant validity.
3.3. Combined data set
Bivariate correlations are shown in Table 7. As expected, both
rm size (Net Sales) and sales growth are highly correlated with
The survey responses were matched with corresponding
nancial performance. Moreover, both ILP and ELP show positive,
secondary data using company names (for those companies that
albeit insignicant, correlations with ROS and statistically sig-
were identied by name in the survey response and for which
nicant positive correlations with sales growth. The relatively
secondary nancial data were available). Descriptive statistics for
high correlation between ILP and ELP is consistent with the notion
secondary nancial data, ELI, and 10 distinct lean practices are
that lean production is a system of practices that should be
presented in Table 5.
Empirical researchers have identied a large number of prac-
tices commonly associated with lean production, which are
Table 6
typically grouped into lean practice bundles. For example, Lean practice bundles (rotated factor pattern).
White and Ruch (1990) identify 10 lean practices and subse-
quently, White et al. (2010) aggregate them into four practice Lean practice Factor 1 external Factor 2 internal
lean practices lean practices
bundles (conformance quality, delivery reliability, volume ex-
ibility, low cost). Similarly, Shah and Ward (2003) categorize 22 Supplier feedback 0.78 0.18
lean practices into four lean bundles (just-in-time, total pro- Supplier JIT 0.70 0.37
ductive maintenance, total quality management, human resource Supplier development 0.70 0.17
management). It can be argued that combining lean practices into Customer involvement 0.63 0.16
Pull system 0.06 0.71
bundles provides parsimony and enhances clarity of exposition.
Continuous ow 0.19 0.70
Setup time reduction 0.39 0.60
Statistical process control 0.34 0.52
Table 5 Employee involvement 0.17 0.76
Descriptive statistics. Total productive maintenance 0.37 0.67

Variable Mean Standard deviation Note: A two factor solution is retained because the eigenvalues of the rst two
factors are greater than 1. The rst two factors explain more than 53% of the
Net sales 12,266 20,003 variation in the data.
Sales growth  0.12 0.14
ROS 0.02 0.19
Total inventory 1,585 2,390
ELI  0.02 0.77
Supplier feedback 4.32 0.67
Supplier JIT 3.51 0.84 Table 7
Supplier development 3.04 0.57 Bivariate correlations.
Customer involvement 3.80 0.59
Pull system 2.96 1.05 Variable 1 2 3 4 5
Continuous ow 3.73 0.69
Setup time reduction 3.26 0.91 1 Net sales
Statistical process control 3.21 0.96 2 Sales growth  0.01
Employee involvement 3.52 0.99 3 Return on sales 0.48 0.30
Total productive maintenance 3.52 0.88 4 Empirical leanness indicator  0.03 0.13  0.04
5 Internal lean practices (ILP) 0.07 0.27 0.11  0.10
Note: Total Assets, Total Inventory, and Net Sales gures are reported in million 6 External lean practices (ELP) 0.15 0.33 0.17 0.14 0.63
US$, Sales Growth and ROS in are reported in percentages and ELI is unitless. All
lean production practices are scored on a scale from 1 (no implementation) to 5 Note: Correlation coefcients printed in bold are statistically signicant at
(complete implementation). po 0.05.
248 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253

implemented simultaneously for optimal performance effects Table 8

(Womack et al., 1990). Mediation tests results for internal and external lean practice bundles.

Independent Dependent variables

Variables
4. Empirical analysis and results A B C D
ROS ELI ROS ROS
4.1. Testing the mediation hypothesis
Intercept  0.19 1.70 0.01  0.03
lnSales 0.00 0.03 0.01 0.00
The mediation hypothesis was tested using the procedure SalesGrowth 0.55nnn 0.44 0.63nnn 0.65nnn
proposed by Baron and Kenny (1986), which is also widely ELP  0.02  0.20  0.02
adopted in operations management literature (e.g., Fullerton ILP 0.08nn  0.40nnn 0.06n
and Wempe, 2009). First, we test the direct effect of lean practice ELI 0.01 0.03
ELI2  0.13nnn  0.12nnn
bundles on nancial performance. We then establish a mediation
path by showing that a lean practice bundle affects inventory n
Signicant at 10%.
leanness which in turn impacts nancial performance. Finally, we nn
Signicant at 5%.
nnn
show that the direct effect of lean production diminishes when Signicant at 1%.
inventory leanness enters the regression model.
Lean production implementation is observed for all rms in
+
the data set (N 229), while ELI and rm performance are only ILP
Financial
observed for public rms that have provided company names Performance
(N 82). In order to use the full sample and obtain unbiased ELP
estimates, we specify a selection model (Greene, 2008, p. 882) as
Lean Production ELI
follows:
Bundles
Z ni b10 b11 LPBi e1i , i 1,2,. . .,229 1
ELI2:
(
1 if Z ni 4 0 Inventory
Zi 2 Leanness
0 if Z ni r 0

FinancialPerf ormancei Fig. 2. Summary of mediation test results.

8 Note: The signs
> b20 b21 lnSalesi b22 SalesGrowthi
< ships, respectively.
b23 LPBi b25 ELIi b26 ELI2i gk e2i if Z i 1
>
:
unobserved if Z i 0
! "  !#
e1i 0 1 rs
N , 4
e2i 0 rs s2
Eq. (1) describes the relationship between a (internal or
external) lean production bundle LPBi for rm i and the likelihood
of observing a rms nancial data Zni , which is a continuous latent
variable. Eq. (2) indicates that a rms nancial data is observed
(Zi 1) when Zni is above a threshold. Eq. (3) expresses a rms
nancial performance conditional upon the observing its nancial
data. Note that inventory leanness ELIi enters this equation in
linear and quadratic terms. As noted previously, we also include
lnSalesi (the natural logarithm of net sales), SalesGrowthi (annual
percentage change in net sales) and gk (industry xed effect) as
control variables (e.g., Rumyantsev and Netessine, 2007). The
error terms e1i and e2i of Eqs. (1) and (3) follow a bivariate normal
distribution specied in Eq. (4). By simultaneously estimating
these equations as a system, one can control for potential
selection bias in the data set. The estimation results are presented
in Table 8.
The results in Table 8 Column A show that internal lean practices
have a signicant and positive effect on nancial performance
(po0.05). External lean practices, however, do not impact nancial
performance. Furthermore, internal lean practices have a signicant
(albeit negative) effect on inventory leanness (Table 8 Column B),
while external lean practices do not signicantly affect inventory
leanness. Consistent with the ndings of Eroglu and Hofer (2011),
the results in Column C show that the relationship between
inventory leanness and nancial performance follows an inverted
U-shape. When lean practice bundles and inventory leanness enter
the regression equation simultaneously (Column D), the effect of
internal lean practices decreases in magnitude and signicance level
(po0.1). Collectively, these ndings support the mediation
and  indicate signicant positive and negative relation-
3
hypothesis H1a for internal lean practices. However, the mediation
hypothesis H1b for external lean practices is not supported. These
results are summarized in Fig. 2.
4.2. Post-hoc analysis of mediation relationships
An unexpected result in Table 8 calls for further exploration.
Specically, the signicant negative effect of internal lean prac-
tices on inventory leanness appears to imply that implementation
of internal lean practices leads rms to hold greater amounts of
inventory. This is not only counterintuitive but it also contradicts
the positive direct effects that internal lean practices have on
nancial performance. Mediation analysis results in which the
direct path and the mediated path exhibit opposite signs are
considered a type of inconsistent mediation (MacKinnon et al.,
2000). Shrout and Bolger (2002) note that the presence of partial
or unexpected mediation effects suggest[s] that the causal
mechanism is more, rather than less, complicated. [y] [T]hese
complications have the potential of enriching both theory and
practice (p. 434). Hence, we further explore the nding of
inconsistent mediation in this post-hoc analysis.
As noted by MacKinnon et al. (2012), timing and longitudinal
effects may be a potential reason for such inconsistent mediation.
Specically, it is conceivable that it may take some time for
internal lean practice implementation to yield the desired bene-
ts in terms of greater inventory leanness. Indeed, there is
evidence that lean production may, at least in some instances,
lead to (temporary) increases in inventory holdings and asso-
ciated costs (Wu, 2002). The investigation of such time-varying
effects, which will require a time series data set, is suggested for
future research.
Mediation analysis further assumes that the causal relationships
between all variables have been properly identied (MacKinnon
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253 249

et al., 2012). Potential simultaneous bidirectional effects may result
in inconsistent estimation results. In this particular case, it is
conceivable that the relationship between lean production and
inventory leanness is bidirectional and simultaneous. In other words,
while lean practices increase a rms inventory leanness, a rm that
is experiencing excess inventory problems (low inventory leanness)
may be more likely to adopt lean practices in an effort to improve its
inventory efciency and nancial performance. Furthermore, internal
and external lean practices may affect each other as well. For
example, a rm that implements internal lean practices may be
more likely to also implement external lean practices and vice versa
(Furlan et al., 2011a, b). Thus, the relationship between lean produc-
tion and inventory leanness may be more than a simple unidirec-
tional relationship.
The central tenet of this post-hoc analysis is that the simulta-
neous bidirectional relationships outlined above may better
capture the true nature of the relationship between lean produc-
tion implementation and inventory leanness and, ultimately,
provide an explanation for the unexpected result shown in
Section 4.1. To capture these complex relationships, a system of
simultaneous linear equations is formulated as follows:
ELIi l11 l12 ILP i l13 ELP i l15 ln Salesi l16 SalesGrowthi e1i ,
5
ILP i l21 l23 ELPi l24 ELIi l25 ln Salesi l26 SalesGrowthi e2i ,
6
ELP i l31 l32 ILPi l34 ELIi l35 ln Salesi l36 SalesGrowthi e3i :
7
In the system of Eqs. (5)(7), inventory leanness is a function of
both internal and external lean practices. At the same time,
internal and external lean practices are functions of inventory
leanness as well as of each other. Table 9 presents estimation
results using two-stage least squares and three-stage least squares
methods which allow for simultaneous estimation of the endogen-
ous relationships between lean practices and inventory leanness.
The estimation results in Table 9 suggest that the implementa-
tion of internal lean practices is positively related to the imple-
mentation of external lean practices, and vice versa. Moreover,
external lean practices and inventory leanness have signicant
positive effects on each other. That is, rms that implement
external lean practices to a greater degree exhibit greater inventory
leanness. Likewise, rms with greater inventory leanness tend to
implement external lean practices to a greater extent. Thus, when
the bidirectional and simultaneous nature of the relationships
between lean practices and inventory leanness is taken into
account, the effect of external lean practices on inventory leanness
is positive and signicant as expected.
Finally, the estimation results from the simultaneous equation
model (Table 9) indicate that internal lean practices have a signi-
cant negative impact on inventory leanness, suggesting that greater
internal lean practice implementation reduces inventory leanness.
However, it should be noted that inventory leanness also has a
signicant negative effect on internal lean practices. In other words,
rms with low inventory leanness, i.e., those that hold above-average
inventories, tend to adopt internal lean practices to a greater extent
than rms with high inventory leanness. This may be an explanation
for the negative coefcient estimate of internal lean practices on
inventory leanness in the mediation analysis results (Table 8).
In summary, the post-hoc analysis achieves two important
objectives: First, it provides more detailed insights into the simul-
taneous and bidirectional relationships between lean practice
bundles as well as between lean production and inventory leanness.
Second, in so doing it offers an explanation for some unexpected
results found in the initial empirical analysis. Specically, the
relationship between ELP and inventory leanness is found to be
positive indeed. Moreover, the post-hoc analysis presents evidence
that the negative relationship between ILP and inventory leanness
may be due to the fact that rms with low inventory leanness tend
to increase their efforts to implement (internal) lean production
practices. These ndings are summarized in Fig. 3.
Internal
Lean Practices
Inventory
Leanness
External
Lean Practices

Table 9
Interrelationships among inventory leanness, internal and external lean practices. Fig. 3. Interrelationships among inventory leanness and lean practices.
Note: The signs and  indicate signicant positive and negative relation-
Independent Dependent variables ships, respectively.

ELI ILP ELP 4.3. Testing the interaction hypotheses

Variables 2SLS 3SLS 2SLS 3SLS 2SLS 3SLS

Intercept  1.46  2.07nnn  0.15  1.05nn 1.79nnn 1.11nnn
The interaction hypothesis posits that when internal and exter-
lnSales  0.07  0.07 0.02  0.02 0.04n 0.03 nal lean practices are implemented together, they will generate a
SalesGrowth 0.22 0.14 0.19 0.09 0.27  0.03 positive interaction effect that will increase the nancial and
ILP  0.46n  0.97nnn 0.47nnn 0.70nnn inventory performance effect of internal and external lean practice
ELP 0.98nnn 1.61nnn 0.93nnn 1.24nnn
bundles. To test the interaction effect, Eq. (8) was estimated.
ELI  0.19n  0.42nnn 0.21nnn 0.41nnn
Perf ormancei b0 b1 lnSalesi b2 SalesGrowthi
Note: 2SLS and 3SLS represent two-stage least squares and three-stage least X
squares methods, respectively. In the 2SLS results, the regression models are all
b3 ILP i b4 ELPi b5 ILP  ELPi gk Indk ei 8
k
signicant at the 5% level and the R2 for ELI, ILP and ELP models are 0.13, 0.39 and
0.45, respectively. Since 3SLS estimates the regression equations as a system, The estimation results are summarized in Table 10 where the
signicance levels are not calculated. In the 3SLS results, the system weighted R2
is 0.61.
dependent variable is nancial performance (ROS, column A) and
n
Signicant at 10%.
inventory leanness (ELI, column B). The respective baseline estima-
nn
Signicant at 5%. tion results without the added interaction effects are shown in
nnn
Signicant at 1%. Table 8. For both ROS and ELI as dependent variables, the interaction
250 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253
Table 10
Moderation tests results for internal and external lean practice bundles.
Independent variables Dependent variables
A ROS
Intercept 0.24
lnSales 0.001
SalesGrowth 0.53nnn
ILP  0.20
ELP  0.29
ILP  ELP 0.08a
a adopt external lean practices. Likewise, we nd evidence that
Marginally signicant in a one-tailed test.
nnn
Signicant at 1%.
effects carry positive and at least marginally signicant coefcient
estimates, suggesting that the performance benets of lean produc-
tion are greater when both practice bundles are implemented
simultaneously. Hence, there is some statistical support for the
interaction hypotheses H2 and H3.
5. Discussion and concluding remarks
Collectively, the analyses presented here draw a more com-
plete picture of the lean production-inventory leanness-nancial
performance triangle. This research adds to the theory of lean
production by highlighting and investigating the mediating role
of inventory management efciency in deriving the nancial
performance benets that are commonly associated with lean
production implementation. As such, this study underscores the
importance of inventory management within the broader realm
of operations management.
5.1. Main ndings
A major nding of this research is the mediating role of
inventories in the relationship between lean practices and rm
performance. Internal lean practices, in particular, have a positive
effect on nancial performance. This direct effect, however,
decreases in magnitude by 25% when inventory leanness enters
the regression equation, thus suggesting that inventory leanness
partially mediates the link between internal lean practices and
nancial performance. This result is consistent with Fullerton and
Wempe (2009) who presented evidence that non-nancial per-
formance measures partially mediate the lean production-nan-
cial performance relationship. This nding further implies that
internal lean practices affect rm performance not only through
improved inventory leanness, but also through other mechan-
isms. Most notably and consistent with prior research internal
lean production practices may directly contribute to greater
nancial performance by lowering operating costs. In this vein,
multiple studies have established that internal lean production
practices such as TQM and TPM are associated with greater
nancial performance (e.g., Cua et al., 2001).
In contrast to internal lean practices, the direct effect of
external lean practices on nancial performance is statistically
insignicant. However, the post-hoc analysis reveals that external
lean practice implementation is positively associated with inven-
tory leanness which, in turn, is linked to nancial performance.
While prior research has found evidence that greater implemen-
tation of external lean practices (e.g., JIT adoption) is associated
with greater nancial performance (e.g., Inman and Mehra, 1993;
Callen et al., 2000; Fullerton et al., 2003) our results indicate that
much of the performance enhancing effect of external lean
production may be due to cost reductions that are derived from
greater inventory leanness.
A second major implication of this study is the identication of
bidirectional and simultaneous effects among internal lean prac-
B ELI tices, external lean practices and inventory leanness. The empiri-
cal results indicate that a rm that implements internal lean
9.41nnn practices is also likely to implement external lean practices. In
0.04
0.29
other words, rms seem to view and implement lean production
 2.75nnn as a comprehensive system instead of a collection of loosely
 2.47nnn related practices that can be individually adopted. In addition,
0.67nnn rms with greater levels of inventory leanness are more likely to
greater external lean practice implementation is associated with
greater inventory leanness. Moreover, we nd evidence that
lower levels of inventory leanness are associated with greater
implementation of internal lean practices. This nding may
indicate that rms recognizing their lack of inventory leanness
implement internal lean practices in an effort to reduce inven-
tories. Hence, this is a possible explanation for the negative
estimate of the effect of internal lean practices on inventory
leanness. Although this nding remains somewhat surprising, it
shows that lean practices vary in their performance effects.
A third major nding of this research is the positive interaction
between internal and external lean practices. From its inception,
lean production was designed as a system of distinct activities
that collectively work to reduce waste and associated costs
(Huang, 1991; Roth and Miller, 1992; Imai, 1998). Our results
support this notion. Specically, we nd that the concurrent
implementation of external and internal lean practices carries
greater performance benets, both in terms of nancial perfor-
mance and inventory leanness, than the implementation of only
one set of lean practices.
5.2. Limitations and future research opportunities
As any research, this study has a number of limitations which
may present interesting future research opportunities. First, the
empirical analyses rely on a cross-sectional data set. The use of
longitudinal data will allow researchers to capture learning
effects in lean production. It is plausible that rms become more
procient in implementing lean practices over time, which may
result in even better nancial performance. Future research could
address this issue by assessing the changes in the effects of lean
practices over time. Second, the negative coefcient estimate for
the effect of internal lean practices on inventory leanness remains
an unresolved issue. Could this nding be a statistical artifact of
the present data set or do internal lean practices really negatively
impact inventory leanness? The replication of this study with a
greater sample size may bring greater clarity to this issue.
5.3. Implications for research and practice
This study has a number of implications for further research.
First, our ndings corroborate the conceptualization of lean
production as a system of management practices that may have
different operational and nancial implications for a rm. The
majority of existing research has focused on the study of JIT
practices and has concluded that greater implementation of such
practices results in lower inventories (e.g., Huson and Nanda,
1995; Balakrishnan et al., 1996). However, there is limited
research examining the relationship between non-JIT lean pro-
duction practices and inventory leanness. Yet, there is theoretical
and anecdotal evidence that lean production may, in at least some
instances, even lead to increases in inventory holdings and
associated costs (Wu, 2002). Moreover, prior research has found
that different lean practices differentially impact non-nancial
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253 251

performance measures (Fullerton and Wempe, 2009). Empirical
research, therefore, must allow for such differential effects. In
addition, as lean production facets are part of a system of
potentially co-dependent practices, the individual effects cannot
be clearly discerned without controlling for potential interactions.
Second, this study underscores the importance of investigating
lean production within the operational context of a rm. As the
results of this study show, inventory leanness mediates, and thereby,
drives the effect of lean production implementation on nancial
performance. As inventory leanness varies signicantly across indus-
tries and rms (Eroglu and Hofer, 2011), it is evident that the nancial
effects of lean production implementation may vary accordingly.
An interesting implication of this research for managers relates
to the positive interaction effect of internal and external lean
practices. In other words, the implementation of a particular lean
practice will not only have a direct performance benet, but it can
also improve the contribution of other existing lean practices in a
rm. Given the synergistic interaction among various lean
practices, managers should look at the system-wide effects of
practices that are adopted in their operations. Also, interactions
between various practices should be assessed to determine the
total performance effect of a given lean practice. Moreover, our
research highlights the importance of considering operational
performance outcomes, such as inventory leanness, as a precursor
to enhanced nancial performance.
Appendix A. Conrmatory factor analysis
See Appendix Table 11.
Appendix B. Empirical leanness indicator (ELI)
The ELI is based on the concept of turnover curves popularized
by Ballou (1981, 2000, 2005) that describes rm size-adjusted

Table 11
Conrmatory factor analysis results for the lean production survey.

External lean practices Supplier feedback Lambda Fit statistics

We frequently are in close contact with our suppliers. 0.6137 w2(8) 9.5238
We give our suppliers feedback on quality and delivery performance. 0.6405 p 0.3000
We strive to establish long-term relationships with our suppliers. 0.7226 CFI 0.9931
Supplier JIT RMSEA 0.0293
Suppliers are directly involved in the new product development process. 0.5648 SRMR 0.0312
Our key suppliers deliver to our plants on JIT basis. 0.5648
We have a formal supplier certication program. 0.3356
Supplier development Lambda Fit statistics
Our suppliers are contractually committed to annual cost reductions. 0.4570 w2(32)41.7159
Our key suppliers are located in close proximity to our plants. 0.3057 p 0.1168
We have corp. level communication on important issues with key suppliers. Dropped CFI 0.9818
We take active steps to reduce the number of suppliers in each category. 0.5068 RMSEA 0.0370
Our key suppliers manage our inventory. 0.5661 SRMR 0.0495
We evaluate suppliers on the basis of total cost and not per unit price. 0.5191
Customer involvement
We frequently are in close contact with our customers. 0.6435
Our customers give us feedback on quality and delivery performance. 0.7505
Our customers are actively involved in current and future product offerings. 0.5829
Our customers are directly involved in current and future product offerings. 0.5504
Our customers frequently share current/future demand info with marketing dept. 0.6875

Internal lean practices Pull Lambda Fit statistics

Production is pulled by the shipment of nished goods. 0.7831 w2(17)26.7066
Production at stations is pulled by the current demand of the next station. 0.7112 p 0.0625
We use a pull production system. 0.9589 CFI 0.9887
We use kanban, squares, or containers of signals for production control. 0.6409 RMSEA 0.0507
Flow SRMR 0.0492
Products are classied into groups with similar processing requirements. 0.5612
Products are classied into groups with similar routing requirements. 0.5774
Equipment is grouped to produce a continuous ow of families of products. 0.7767
Families of products determine our factory layout. 0.6473
Setup Lambda Fit statistics
Our employees practice setups to reduce the time required. 0.8206 w2(8) 14.9280
We are working to reduce the setup times in our plants. 0.6568 p 0.0606
We have low setup times in our plants. 0.5258 CFI 0.9850
SPC RMSEA 0.0625
Large number of equipment/processes on shop oor are currently under SPC. 0.7946 SRMR 0.0448
We make extensive use of statistical techniques to reduce process variance. 0.9481
Charts showing defect rates are used as tools on the shop oor. 0.6655
We use shbone type diagrams to identify causes of quality problems. Dropped
We conduct process capability studies before product launch. Dropped
Employee involvement Lambda Fit statistics
Shop-oor employees are key to problem solving teams. 0.8001 w2(8) 9.4532
Shop-oor employees drive suggestion programs. 0.8219 p 0.3055
Shop-oor employees lead product/process improvement efforts. 0.8000 CFI 0.9974
Shop-oor employees undergo cross-functional training. Dropped RMSEA 0.0286
TPM SRMR 0.0247
We dedicate a portion of everyday to planned equipment maintenance related activities. 0.6037
We maintain all our equipment regularly. 0.8416
We maintain excellent records of all equipment maintenance related activities. 0.8011
We post equipment maintenance records on shop oor for active sharing with employees. Dropped

Note: All lambda values are signicant at p o 0.01 level.

252 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253
Firm Size-Adjusted
Industry Average
ELI
Inventory
Sales
Fig. 4. The salesinventory relationship. Note: Figure adapted from Eroglu and
Hofer (2011).
industry average inventory levels (Fig. 4). A rms deviation from
this curve represents its level of leanness or unleanness.
More formally, ELI is calculated as the error term from the
regression model lnInvif a0i a1i lnSalesif eif , where Invif is the
average of the rms total inventories reported at the end of 2008
and 2009, and Salesif is the total sales volume of rm f in industry i.
Fitting this model by industry provides some control for industry-
specic idiosyncrasies of the inventory-sales relationship. The
remaining differences in the inventory-sales relationship across
rms within an industry, thus, primarily arise from systematic
differences in rms inventory management practices. The residual
eif is studentized and rescaled such that relatively lower inventory
holdings translate to higher positive ELI values, and vice versa
(Eroglu and Hofer, 2011). The model shown above was estimated
using the ordinary least squares method for each of the 24
four-digit NAICS industries represented in the survey sample.
Detailed regression results are not reported here due to space
constraints but are available from the authors upon request. The
average value of the coefcient estimate of the lnSales variable, an
indicator of scale economies in inventory management, was 0.91.
The average value of the R-squared statistic was 0.89, indicating a
very good model t.
References
Armstrong, J.S., Overton, T.S., 1977. Estimating non-response bias in mail surveys.
Journal of Marketing Research 14 (3), 396402.
Balakrishnan, R., Linsmeier, T.J., Venkatachalam, M., 1996. Financial benets from
JIT adoption: effects of customer concentration and cost structure. The
Accounting Review 71 (2), 183205.
Ballou, R.H., 1981. Estimating and auditing aggregate inventory levels at multiple
stocking points. Journal of Operations Management 1 (3), 143153.
Ballou, R.H., 2000. Evaluating inventory management performance using a turn-
over curve. International Journal of Physical Distribution & Logistics Manage-
ment 30 (1), 7285.
Ballou, R.H., 2005. Expressing inventory control policy in the turnover curve.
Journal of Business Logistics 26 (2), 143164.
Barber, B.M., Lyon, J.D., 1996. Detecting abnormal operating performance: the
empirical power and specication of test statistics. Journal of Financial
Economics 41, 359399.
Bardhan, I., Mithas, S., Lin, S., 2007. Performance impacts of strategy, information
technology applications, and business process outsourcing in US manufactur-
ing plants. Production and Operations Management 16 (6), 747762.
Baron, R.M., Kenny, D.A., 1986. The moderatormediator variable distinction in
social psychological research: conceptual, strategic, and statistical considera-
tions. Journal of Personality and Social Psychology 51 (6), 11731182.
Biggart, T.B., 1997. The effects of just-in-time inventory system adoption on rm
performance. Dissertation. The Florida State University, USA.
Biggart, T.B., Gargeya, V.B., 2002. Impact of JIT on inventory to sales ratios.
Industrial Management Data Systems 102 (3/4), 197202.
Braunscheidel, M.J., Suresh, N.C., 2009. The organizational antecedents of a rms
supply chain agility for risk mitigation and response. Journal of Operations
Management 27 (2), 119140.
Callen, J.L., Fader, C., Krinsky, I., 2000. Just-in-time: a cross-sectional plant analysis.
International Journal of Production Economics 63 (3), 277301.
Cannon, A.R., 2008. Inventory improvement and nancial performance. Interna-
tional Journal of Production Economics 115 (2), 581593.
Capkun, V., Hameri, A., Weiss, L.A., 2009. On the relationship between inventory
and nancial performance in manufacturing companies. International Journal
of Operations & Production Management 29 (8), 789806.
Chen, H., Frank, M.Z., Wu, O.Q., 2005. What actually happened to the inventories of
American companies between 1981 and 2000? Management Science 51 (7),
10151031.
Chen, H., Frank, M.Z., Wu, O.Q., 2007. US retail and wholesale inventory perfor-
mance from 1981 to 2004. Manufacturing & Service Operations Management 9
(4), 430456.
Chikan, A., 2009. An empirical analysis of managerial approaches to the role of
inventories. International Journal of Production Economics 118 (1), 131135.
Chikan, A., 2011. Managers view of a new inventory paradigm. International
Journal of Production Economics 133 (1), 5459.
Claycomb, C., Droge, C., Germain, R., 1999. The effect of just-in-time with
customers on organizational design and performance. International Journal
of Logistics Management 10 (1), 3758.
Claycomb, C., Germain, R., Droge, C., 1999. Total system JIT outcomes: inventory,
organization and nancial effects. International Journal of Physical Distribu-
tion & Logistics Management 29 (10), 612630.
Cua, K.O., McKone, K.E., Schroeder, R.G., 2001. Relationships between implementa-
tion of TQM, JIT, and TPM and manufacturing performance. Journal of
Operations Management 19 (6), 675694.
Demeter, K., Matyusz, Z., 2011. The impact of lean practices on inventory turnover.
International Journal of Production Economics 133 (1), 154163.
Dobos, I., 2007. Tradable permits and production-inventory strategies of the rm.
International Journal of Production Economics 108 (12), 329333.
Droge, C., Germain, R., 1998. The just-in-time inventory effect: does it hold under
different contextual, environmental, and organizational conditions? Journal of
Business Logistics 19 (2), 5371.
Eroglu, C., Hofer, C., 2011. Lean, leaner, too lean? The inventory-performance link
revisited. Journal of Operations Management 29 (4), 356369.
Evers, P.T., 1995. Expanding the square root law: an analysis of both safety and
cycle stocks. Transportation Research Part E: Logistics and Transportation
Review 31 (1), 120.
Faireld, P.M., Yohn, T.L., 2001. Using asset turnover and prot margin to forecast
changes in protability. Review of Accounting Studies 6, 371385.
Fullerton, R.R., McWatters, C.S., 2001. The production performance benets from
JIT implementation. Journal of Operations Management 19 (1), 8196.
Fullerton, R.R., McWatters, C.S., Fawson, C., 2003. An examination of the relation-
ships between JIT and nancial performance. Journal of Operations Manage-
ment 21 (4), 383404.
Fullerton, R.R., Wempe, W.F., 2009. Lean manufacturing, non-nancial perfor-
mance measures, and nancial performance. International Journal of Opera-
tions & Production Management 29 (3), 214240.
Furlan, A., Dal Pont, G., Vinelli, A., 2011a. On the complementarity between
internal and external just-in-time bundles to build and sustain high perfor-
mance manufacturing. International Journal of Production Economics 133 (2),
489495.
Furlan, A., Vinelli, A., Dal Pont, G., 2011b. Complementarity and lean manufactur-
ing bundles: an empirical analysis. International Journal of Operations and
Production Management 31 (8), 835850.
Germain, R., Droge, C., Spears, N., 1996. The implications of just-in-time for
logistics organization management and performance. Journal of Business
Logistics 17 (2), 1934.
Goldsby, T.J., Grifs, S.E., Roath, A.S., 2006. Modeling lean, Agile, And Leagile
supply chain strategies. Journal of Business Logistics 27 (1), 5781.
Greene, W.H., 2008. Econometric Analysis. Pearson Prentice Hall, Upper Saddle
River, NJ.
Guinipero, L.C., Pillai, K.G., Chapman, S.N., Clark, R.A., 2005. A longitudinal
examination of JIT purchasing practices. International Journal of Logistics
Management 16 (1), 5170.
Huang, P., 1991. World class manufacturing in the 1990s: integrating JIT, TQC, FA,
and TPM with worker participation. In: Fandel, G., Zapfel, G. (Eds.), Modern
Production Concepts: Theory and Applications. Springer, New York,
pp. 491507.
Hult, G.T., Ketchen, D.J., Arrfelt, M., 2007. Strategic supply chain management:
improving performance through a culture of competitiveness and knowledge
management. Strategic Management Journal 28 (10), 10351052.
Huson, M., Nanda, D., 1995. The impact of just-in-time manufacturing on rm
performance in the US. Journal of Operations Management 12 (3/4), 297310.
Imai, M., 1998. Will Americas corporate theme song be just in time? Journal of
Quality Participation 21 (2), 2628.
Inman, R.A., Mehra, S., 1993. Financial justication of JIT implementation. Inter-
national Journal of Operations & Production Management 13 (4), 3239.
Irvine, F., 2003. Long term trends in US inventory to sales ratios. International
Journal of Production Economics 8182, 2739.
Jayaram, J., Vickery, S., Droge, C., 2008. Relationship building, lean strategy and
rm performance: an exploratory study in the automotive supplier industry.
International Journal of Production Research 46 (20), 56335649.
Kinney, M.R., Wempe, W.F., 2002. Further evidence on the extent and origins of
JITs protability effects. The Accounting Review 77 (1), 203225.
Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling. The
Guildford Press, New York.
Konecny, P.A., Thun, J., 2011. Do it separately or simultaneously an empirical
analysis of conjoint implementation of TQM and TPM of plant performance.
International Journal of Production Economics 133 (2), 496507.
 C. Hofer et al. / Int. J. Production Economics 138 (2012) 242253
Koumanakos, D.P., 2008. The effect of inventory management on rm perfor-
mance. International Journal of Productivity and Performance Management 57
(5), 355369.
Lambert, D.M., Harrington, T.C., 1990. Measuring non-response bias in mail
surveys. Journal of Business Logistics 11 (2), 525.
Matsui, Y., 2007. An empirical analysis of just-in-time production in Japanese
manufacturing companies. International Journal of Production Economics 108
(12), 153164.
MacKinnon, D.P., Coxe, S., Baraldi, A.N., 2012. Guidelines for the investigation of
mediating variables in business research. Journal of Business and Psychology
27, 114.
MacKinnon, D.P., Krull, J.L., Lockwood, C.M., 2000. Equivalence of the mediation,
confounding and suppression effect. Prevention Science 1 (4), 173181.
Mistry, J.J., 2005. Origins of protability through JIT processes in the supply chain.
Industrial Management Data Systems 106 (6), 752768.
Miyazaki, S., 1996. An analytical comparison of inventory costs between the pull
and the parts-oriented production systems. International Journal of Produc-
tion Economics 44 (12), 151157.
Norris, D.M., Swanson, R.D., Chu, Y., 1994. Just-in-time production systems: a
survey of managers. Production and Inventory Management Journal 35 (2),
6366.
Obermaier, R., 2012. German inventory to sales ratios 19712005an empirical
analysis of business practice. International Journal of Production Economics
135, 964976.
Podsakoff, P.M., MacKenzie, S.B., Lee, J., Podsakoff, N., 2003. Common method
biases in behavioral research: a critical review of the literature and recom-
mended remedies. Journal of Applied Psychology 88 (5), 879903.
Relph, G., Barrar, P., 2003. Overage inventory-how does it occur and why is it
important? International Journal of Production Economics 8182, 163171.
Roth, A.V., Miller, J.G., 1992. Success factors in manufacturing. Business Horizons
35 (4), 7381.
Rumyantsev, S., Netessine, S., 2007. What can be learned from classical inventory
models? A cross-industry exploratory investigation. Manufacturing & Service
Operations Management 9 (4), 409429.
253
Shah, R., Ward, P.T., 2003. Lean manufacturing: context, practice bundles, and
performance. Journal of Operations Management 21 (2), 129149.
Shah, R., Ward, P.T., 2007. Dening and developing measures of lean production.
Journal of Operations Management 25 (4), 785805.
Shook, C.L., Ketchen, D.J., Hult, G.T., 2004. An assessment of the use of structural
equation modeling in strategic management research. Strategic Management
Journal 25 (4), 397404.
Shrout, P.E., Bolger, N., 2002. Mediation in experimental and nonexperimental studies:
new procedures and recommendations. Psychological Methods 7 (4), 422445.
Stock, J.R., Broadus, C.J., 2006. Doctoral research in supply chain management
and/or logistics-related areas: 19992004. Journal of Business Logistics 27 (1),
139142.
Swamidass, P.M., 2007. The effect of TPS on US manufacturing during 19811998:
inventory increased or decreased as a function of plant performance. Interna-
tional Journal of Production Research 45 (16), 37633778.
White, R.E., 1993. An empirical assessment of JIT in US manufacturers. Production
and Inventory Management Journal 34 (1), 3842.
White, R.E., Ruch, W.A., 1990. The composition and scope of JIT. Operations
Management Review 7 (3/4), 918.
White, R.E., Ohja, D., Kuo, C.C., 2010. A competitive progression perspective of JIT
systems: evidence from early US implementations. International Journal of
Production Research 48 (2), 61036124.
Wiseman, R.M., 2009. On the use and misuse of ratios in strategic management
research. In: Bergh, D.D., Ketchen, D.J. (Eds.), Research Methodology in
Strategy and Management, 5. JAI Press, Bingley, UK, pp. 75110.
Womack, J.P., Jones, D.T., Roos, D., 1990. The Machine that Changed the World.
Rawson Associates, New York.
Womack, J.P., Jones, D.T., 1996. Beyond Toyota: how to root out waste and pursue
perfection. Harvard Business Review 74 (5), 140158.
Wu, Y.J., 2002. Effective lean logistics strategy for the auto industry. The
International Journal of Logistics Management 13 (2), 1938.
Yang, M.G., Hong, P., Modi, S.B., 2011. Impact of lean manufacturing and environ-
mental management on business performance: an empirical study of manufac-
turing rms. International Journal of Production Economics 129 (2), 251261.