ISO 14001 and Financial Performance (2008) 21pp

7/29/2019 ISO 14001 and Financial Performance (2008) 21pp

http://slidepdf.com/reader/full/iso-14001-and-financial-performance-2008-21pp 1/21

210 Int. J. Business Excellence, Vol. 1, Nos. 1/2, 2008

Copyright © 2008 Inderscience Enterprises Ltd.

The effect of the ISO-14001 environmentalmanagement system on corporatefinancial performance

Jiangning Zhao

Department of Management

Charlton College of Business

University of Massachusetts Dartmouth

Dartmouth, MA, 02747–2300, USA

E-mail: [email protected]

Abstract: This study employed a quasi-experimental design to compare ISOcompanies with the matched comparison companies on the effect of ISO-14001registration (independent variable) on financial performance (three dependentvariables): Return On Revenue (ROR), Return On Assets (ROA), andOperating Revenue (OPR).

Eighty-one pairs of companies (N = 162) were collected from the Standard& Poor’s (S&P) database, which had at least one ISO certificate and for whomfive years of financial data were available. Data on ISO registration indicated astrong pattern of distribution by industry, by years, and by the size of company.

Control for initial difference, the results of statistical analyses (ANCOVA,MANOVA, and ANOVA) indicated that registration of ISO-14001 led tolower ROR and ROA in the two-year period, and no change in the thirdyear period. OPR did not change. Early adoption, length of ISO registration andthe size of company did not affect the relationship of ISO registration and

financial performance.

Keywords: ISO 14001; Environmental Management System; EMS; impact;financial performance; dynamic capabilities; triple bottom line.

Reference to this paper should be made as follows: Zhao, J. (2008) ‘The effectof the ISO-14001 environmental management system on corporate financialperformance’, Int. J. Business Excellence, Vol. 1, Nos. 1/2, pp.210–230.

Biographical notes: Jiangning Zhao a Senior Researcher on eco-environmentand natural resources, had worked for Chinese Academy of Science for 14years before he received his MBA and PhD in management fromMaharishi University of Management. His current research interest is on thestrategic dynamisms of environmental management system, throughout thecontinuing process of system design, implementation, monitoring, assessingand system modification.

1 Introduction

Since it was initiated in 1996, ISO-14001 has drawn worldwide attention. Over 103 583

companies have registered as of January 2006 (Corporate Risk Management, 2006) with

various motivations and expectations. However, one puzzle still remains unsolved. That



The effect of the ISO-14001 environmental management system 211

is: does ISO-14001 contribute to financial performance? Providing an answer to this

question could promote the direct motivation for companies to register. Furthermore,

providing an answer to this question has broad-reaching implications not only for

companies who wish to be green, but also for the sustainability of ecosystem as well.

1.1 Research question

Delmas (2006) stipulated that companies will seek ISO-14001 certification only when

it is in their best economic interest, i.e., the benefits of adopting ISO-14001 certification

outweigh the costs. However, the effect of ISO-14001 on financial performance

has not previously been studied empirically. To fill the need in the existing literature, this

study is designed to examine if the registration of ISO-14001 is related to corporate

financial performance.

Employing a quasi-experimental design, this study statistically examined sixhypotheses. The first three hypotheses were designed to test the effect of ISO-14001

EMS registration on the three financial variables: Return on Revenue (ROR), Return On

Assets (ROA), and Operating Revenue (OPR). The other three hypotheses were designed

to test the function of three factors that might influence the effect of ISO-14001

registration on financial performance: (1) early adoption, (2) path-dependent learning

(length of registration) and (3) size of company.

1.2 Theoretical frameworks

1.2.1 Triple-Bottom-Line (TBL)

TBL, the ultimate goal of ISO-14001 EMS (Edwards, 2004; von Zharen, 2001), is an

organisational strategic philosophy aimed at broadening organisational goals froma sole focus on profit-making bottom-line to a three-fold bottom-line including an

environmental and a social responsibility bottom-line. Since the mid-1990s, the TBL

philosophy has drawn widespread attention related to a broad range of management

areas, including organisational strategy (Starik and Marcus, 2000), organisational

environmental performance (Toffel, 2003), environmental leadership (Egri and Herman,

2000), organisational sustainability (Stead et al., 2004; Wilson and Lombardi, 2001),

marketing strategy (Geffen and Rothenberg, 2000) and environmental accounting

(Parker, 2000).

If ISO-14001 is effective in improving both the environmental-bottom-line and the

financial-bottom-line simultaneously, then TBL is actionably achievable by the adoption

and implementation of ISO-14001, since the social-bottom-line is an integrated

manifestation of the other two lines.

1.2.2 Dynamic capabilities

Dynamic capabilities were defined by Teece et al. (1997, p.515) as “The firms’ ability to

integrate, build, and reconfigure internal and external competencies to address rapidly

changing environments”. In Russo’s (2007) study on the environmental performance of

ISO-14001, dynamic capabilities were defined and examined as the combination of three

elements: early timing strategy, path-dependent experience, and organisational ability of

integration. The present study examines two of these same dynamic capabilities: early

timing, and path dependent experience. If this study can detect a difference between



212 J. Zhao

companies that adopted ISO-14001 EMS earlier and companies that adopted later, then

this element is fit to explain not only the effect of adoption of ISO-14001 onenvironmental performance (King and Lenox, 2002; King et al., 2005; Russo, 2007) but

also the effect on financial performance.

The same test applies to the effect of path-dependent experience, which is

operationalised by Russo (2007) as length of registration; it can be used to detect

differences between companies that have longer periods of registration and companies

with shorter periods of registration. The third element of dynamic capabilities, the ability

of integration, was not directly tested in this study. It can be induced or extrapolated from

the results of the other two elements. If differences exist in both of the two tested

elements, then it is reasonable to conclude that the theory of dynamic capabilities is

relevant to ISO-14001 EMS.

2 ISO-14001 literature

2.1 An overview of ISO-14001 literature

ISO-14001 is an impact-oriented Environmental Management System (EMS). The

revised version of ISO-14001 provides operational guidelines for registered companies to

implement, measure, monitor, and control an organisation’s activities and their impacts.

The dynamic cycle of implementation (Plan-Do-Check-Act), once initiated, must be a

continuing process throughout the entire management system including resource input,

throughput, and output. Impacts of activities must be precisely monitored, measured and

evaluated (IEMA, 2005).

Successful implementation of ISO-14001 EMS have been anecdotally empirically

reported that implementation improves environmental performance such as reducingwaste, emissions and pollution; fostering energy reuse; and therefore resulting in cost

savings (Christmann and Taylor, 2003; King and Lenox, 2002; King et al., 2005;

Russo, 2002; Russo, 2007; Russo and Harrison, 2005). Theoretically, ISO-14001

has been predicted as a system capable of leading organisations towards the goal of

triple-bottom-line (Edwards, 2004; Stead et al., 2004; von Zharen, 2001). These studies

have delineated a roadmap from adoption, through implementation, and to the impacts.

However, since ISO-14001 is a voluntary process, and does not bring tangible values

into registered companies (Delmas, 2006), it is critically important for ISO companies to

directly signal to stakeholders that the management of a certified company is

environmentally sound and financially profitable and tangible.

2.2 Theoretical issues in ISO-14001 literature

The review of ISO-14001 literature can be theoretically categorised into three groups: the

impact of ISO-14001 on environmental performance, the impact of ISO-14001 on

organisational strategic management, and the impact of ISO-14001 on the ‘costs and

benefits’. In general, ISO-14001 research is dominated by the relationship between the

registration of ISO-14001 and environmental performance (see Table 1).




Table 1 Theories applied by researchers in ISO-14001 EMS literature

Theory Theme Major authors

Impact of environmentalperformance

Impact of ISO-14001 registration onenvironmental performance (emissionsreduction and waste reduction)

Impact of ISO-14001 registration onorganisational performance(effectiveness and qualityimprovement in operation)

King and Lenox (2002); Melnyket al. (2002); Russo (2002); Russoand Harrison (2005)

Organisationalstrategicmanagement

The impact of ISO-14001 registrationon organisational change (institutionand opportunities)

The impact of ISO-14001 registration

on organisational competence(dynamic capabilities)

Jiang and Bansal (2003); Kinget al. (2005); Russo (2007)

Does it pay tobe green?

Relationship between environmentalperformance (emission reduction,waste reduction, and risk reduction)and financial performance(cost savings)

Bansal and Bogner (2002); Delmas(2006); Hamschmidt and Dyllick(2001); King and Lenox (2002)

Russo (2007) suggested that a solid theory is needed to explain the nature and function

of ISO-14001 EMS. However, the theory that ISO-14001 leads to improved

environmental performance has not been completely verified yet. For example, in their

recent study, King et al. (2005) concluded that ISO-14001 EMS is not associated with

improved environmental performance. It indicated that the linear relation from ISO

registration to environmental performance, and to better financial performance need to befurther studied.

The theory of dynamic capabilities refers to change-oriented organisational and

strategic routines (early timing and path dependent experience) by which managers are

empowered to structure and restructure, organise and reorganise organisational resources

in order to generate value-adding capabilities (Russo, 2007). Evidence found that early

adoption of ISO-14001 relates to environmental impacts (Russo, 2007). However, the

theory of dynamic capabilities has not been tested in the effect of adoption of ISO-14001

on financial performance. Therefore, the review of ISO-14001 literature provides a

theoretical ground for the present study: are organisations capable of integrating the

dynamic capabilities into their ongoing system to improve financial performance?

2.3 Methodological issues in ISO-14001 literature

Methodologically, Toxic Release Inventory (TRI) has been used by most of researches on

ISO-14001 as the measure of environmental performance of ISO-14001 (see Table 2),

indicating a conceptual poverty in measurement – without comparison group, whether

TRI is attributed to the effect of EMS remains unclear.

To date, none of the studies in ISO-14001 literature employed quasi-experimental

design using a comparison group (see Table 2), a history threat to the internal validity

might be the major drawback of the existing literature.



214 J. Zhao

Table 2 Measures applied in ISO-14001 literature

ISO-14001

Author year EMS indicator

Environmentalmeasures Research design

Research findings

King et al.(2005)

ISO-14001registration

TRI Lowerenvironmentalperformanceis related toregistration

Russo (2007) ISO-14001registration

TRI

Longitudinal

Non-Quasi-Experimentaldesign

No matched control group

No control forinitial difference

Regression analysis

Positive butweak

The theoretical and methodological poverty of ISO-14001 literature provides this studywith an opportunity to verify if ISO-14001 is a TBL – oriented management system, not

only environmental incentives but also economical incentives.

2.4 Research hypotheses

Six hypotheses have been established. The first three are to examine the effect of

adoption and implementation of ISO-14001 on financial performance. Hypotheses 4, 5

and 6 are dedicated to the effect of dynamic capabilities.

Hypothesis 1 Companies that adopt ISO-14001 EMS gain different ROR compared

to matched companies that do not adopt ISO-14001

Hypothesis 2 Companies that adopt ISO-14001 EMS gain different ROA compared


Hypothesis 3 Companies that adopt ISO-14001 EMS gain different OPR compared


2.4.1 Early adoption of ISO-14001

Russo (2007) suggested that an important element of dynamic capabilities is the ability to

move quickly when new opportunities to improve appear, and speedy initial movement

would suggest that the organisation is capable of appreciating external opportunities.Therefore, it is reasonable to assume that early adoption may provide organisations with a

more favourable ratio of costs and benefits. Thus:

Hypothesis 4 Early adopters of ISO-14001 EMS gain different financial

performances than those that adopt later




2.4.2 Path dependent experience

The process of ISO-14001 EMS is formalised by the ‘Plan-Do-Check-Act’ cycle

(International Organization for Standardization, 2002), and then normalised into

organisational routines. This cyclical process functions as a dynamic progressing path for

registrants to accumulate experiences that propel them to progress. Thus:

Hypothesis 5 Companies that have a longer period of certification with ISO-14001

EMS experience different financial performance than those companies

with a shorter period of certification

2.4.3 Size of company

One of the functions of ISO-14001 EMS is to increase the ability of certified facilitiesto efficiently internalise external resources (von Zharen, 2001). This ability is

related to parent company’s size and the size of facility itself (King et al., 2005;

Russo, 2007). Limited by the data of size on the facilities’ level, this study investigates a

sixth hypothesis:

Hypothesis 6 Larger companies experience different financial performance from

ISO-14001 EMS than smaller companies

All the six hypotheses in the present study are non-directional, suggesting that the impact

of ISO-14001 registration and implementation on financial performance could be an

increase or a decrease. Although increases in financial performance are expected by some

advocates of ISO-14001, the lack of any findings leads to undertake this exploratory with

an openness to if there any impact, positive or negative.

3 Research design and methods

Methodologically, this study employed a quasi-experimental design to compare the

experimental group (ISO companies) with the matched comparison group (non-ISO

companies). By examining companies in a wide range of industries (44 industries, SIC

codes: 1000–9900), this study extended the scope of generalisability. It may involved a

risk of dispersed sample, where different industry environment influences comparability.

However, by employing a matched comparison group (matched by the same years, same

industry, and similar size of average total assets), this potential confound to the validity

of statistical conclusion was controlled.

3.1 Study period

It was reported that ISO-9000 implementation takes about 6 to 18 months (Corbett et al.,

2005). Since the nature of ISO-14001 is similar to ISO-9000, it is reasonable to assume

that the same period of implementation applies to ISO-14001. Given this assumption, this

study averaged two or three years’ scores before and after registration as pre- and

post- scores of the dependent variable (T – 2, or T – 3; and T + 2, or T + 3). T stands for

the fiscal reporting year during which the first certification was received. Financial data



216 J. Zhao

at year T (registration year) was not counted in pre- or post-values. T – 1 is the year prior

to the registration, T + 1 is the year after the registration, same as T – 2, T – 3, T + 2, andT + 3 (see Figure 1).

Figure 1 Fiscal reporting years in the study period

The study period (T ± 2 or T ± 3) was determined according to the date of initial

certification; every company selected (both ISO and non-ISO companies) was required to

have financial data in the (T ± 2) period. All the companies in the (T ± 3) period were

selected from the pool of (T ± 2). This method of defining the actual effect of ISO-14001

registration was modelled on the design of the National Database on Environmental

Management System (NDEMS Final Report, 2003) and the study on financial impact of

ISO-9000 (Corbett et al., 2005).

In order to include two or three years of reported financial performance after year

T , the study was limited to companies who received certification no later than 2001

(for T ± 2 analyses), and no later than 2002 (for T ± 3 analyses).

3.2 Data and sample

3.2.1 Registration data

Data on ISO registration was collected from Quality System Update Publishing Company

(QSU) and six accredited ISO registrars in North America. According to updated

requirements (ISO, 2004), all registered companies must be regularly monitored and

inspected using a standard process of documentation, communication, and evaluation by

registrars. Therefore, there is a solid base for the reliability of the registration data, which

were sampled in three phases, as outlined in Figure 2.

In the first phase, this study matched 5106 ISO companies with 1279 usable

companies in Standard & Poor’s NetAdvantage Financial Database Room (2005)

according to the name and certificate. This matching procedure provided 214 companies

that are both ISO and Standard & Poor (S&P) companies.

In the second phase, the sample was reduced to those companies which had financialdata on the three dependent variables for two years (±2) before and after the fiscal

year in which the first certificate was obtained (year T); and which had a matched

comparison company from the same industry, same period of time, and within the same

capitalisation level. This matching procedure ended with 90 pairs of companies

(N = 180), a yield of 42.06% for the (T ± 2) periods, only 52 out of 90 pairs of companies

(N = 104) for the (T ± 3) periods.




Figure 2 Data sources and sampling procedure

In the third phase, the sample of 90 was further reduced to 81 ISO companies (N = 162)

that had matching companies with assets’ size between 30% and 300% of the assets’ size

of the ISO Company.1 Of these 81, 45 pairs of companies (N = 90) had data for the time

period (T ± 3).

3.2.2 Matching procedure

The matching process was as follows. A random number was used to select a matched

non-ISO company for the ISO Company with earliest certification; then, in order to

ensure the same population of selecting matched comparison companies, this selected

matching company was put back into the pool of non-ISO companies for selection

of the next match. If the same matching company was selected twice, then it was put back

into the pool again and another selection was made until a new company had been

randomly selected. Using this method of matching keeps the sample population relatively



218 J. Zhao

consistent, therefore, strengthens the generalisability. If an ISO company remained

unmatched because of the shortage of non-ISO companies, then that ISO Company mustbe dropped out.

A company could be considered in the non-ISO matching pool for a particular ISO

company even if the non-ISO company eventually did get ISO certification as long as

that registration was not during the periods of (T ± 2, or T ± 3) for which financial data

was analysed. For example, Ford registered in 1996 and General Motors registered in

1999; therefore, GM was qualified as a matching company for the period (T ± 2).

3.2.3 Financial data

Financial data were selected from Industry Analysis of the Standard & Poor’s

NetAdvantage Financial Database Room (2005), which includes S&P 500 large cap, S&P

400 mid cap, and S&P 600 small cap companies. In order to be selected, (1) all the ISO

companies must have financial data on ROR, ROA and OPR in at least two consecutive

years before and after ISO registration; (2) each of the ISO companies must have a

matched comparison company with all the financial data for the same years.

3.3 Measures

3.3.1 Dependent variables

ISO-14001 registration is issued at the facility level, not at the company level; however,

financial data was limited to the company’s level. This situation makes the results of tests

more conservative because when the sample “contains many multi-site firms with only

one certified site, any effects of certification must be strong in order to be detected”

(Corbett et al., 2005, p.1049). However, there is a risk that the present study might obtain

a ‘diluted effect’ (discussed later).In addition to the ratios of financial return (ROR and ROA), OPR is an important

factor in evaluating the financial performance resulting from systematic operation

(Brigham, 1995). Therefore, this study decided to use ROR and ROA and OPR as

three measures of financial performance. All the three measures were properly defined by

S&P document.

3.3.2 Independent variables

To test Hypotheses 1–3, this study compared the ISO companies registered during the

period of 1996 to 2001 for (T ± 2), or the period of 1996 to 2002 for (T ± 3), with

matched comparison companies for the same period.

To test Hypothesis 4, this study used the data of registration year (1996/1997) as an

independent variable for early adoption. According to Russo (2007), the demarcation of

early and late adoptions of ISO-14001 is defined as the end of 1997, after which the

standard was well-established and spreading globally.

To test Hypothesis 5, this study used the actual number of registration years as an

independent variable. Registration date was defined as the date of issuance of the first

certificate from the company to which the certified facility belongs. Years of registration

was measured based on the company’s fiscal reporting year and actual date of

registration: if the fiscal reporting year ended six months prior to the date of registration,




then it was counted as a full year, otherwise it was rolled over to the next year. As for

dependent variables, this study used post-scores of ROA, ROR, and OPR in year 2003 to

examine if there were any changes attributable to the numbers of years of registration.

To test Hypothesis 6, the effect of company size, this study used Average Total

Assets (ATA) as the measure of company size. This measure was introduced by Corbett

et al. (2005) in their study of the financial impact of ISO-9000. ATA is calculated as Net

Income divided by ROA. Data on net income and ROA was drawn from the Standard &

Poor’s NetAdvantage Financial Database Room (2005).

3.4 Statistical analysis

Analysis of covariance (ANCOVA) was chosen as the primary tool for statistical analysis

because of the dual function of a covariate – It increases statistical power and

simultaneously increases strength of research control for the initial difference on thedependent variables (Huck, 2004; Stevens, 1997). Multivariate Analyses of Variance

(MANOVA) and Analysis of Variance (ANOVA) were used when the assumptions of

ANCOVA could not be met (Huck, 2004). All the statistical analyses were conducted

using log-transformed data (both pre- and post-scores of the three dependent variables) to

avoid skewness of data distribution.

In order to examine if the difference on the pre-scores of the three dependent

variables was significantly strong enough to cause the change on the post-scores, and if

the initial scores of asset sizes of the ISO companies were properly matched with

matched comparison companies, examinations were conducted using T -test, and the

results showed that the differences on pre-scores of ISO companies and their matched

comparison companies are not significant for the three dependent variables, neither the

initial scores of asset size in the study period (T < T

*

, P > 0.5, see Table 3).

Table 3 T -test on the equality of pretest variables

T ± 2 (N = 162) T ± 3 (N = 90)

T-value P-value Decision T-value P-value Decision

DV T ± 2 (T*=2.4187) T ± 3 (T* = 2.4398)

ROR-Pre 0.1239 0.901578 Accept 0.5152 0.607672 Accept

ROA-Pre 0.1195 0.905033 Accept 0.5494 0.584137 Accept

OPR-Pre –0.4069 0.684659 Accept –0.0120 0.990466 Accept

Asset size pre –0.0254 0.97982 Accept 0.2761 0.783748 Accept

To further control for research validity, this study used the ROA baseline value as thesecond covariate in addition to the pre-scores of the dependent variables. The rationale is

that more covariates can substantially reduce the “inter-correlations among the covariate

variables by reducing different part of error variance from the dependent variable(s)”

(Stevens, 1997, p.332). Therefore, by adding the second covariate variable, this study

made a better adjustment for initial differences among the two groups (the ISO group and

non-ISO group). In addition, since there were three dependent variables, the chances of

Type I errors are inflated when separate ANCOVAs are used. To control for this

statistical bias, this study adopted the Bonferroni Adjustment (α = 0.05/3 = 0.0167).



220 J. Zhao

3.5 Tests of statistical assumptions

To ensure the validity and reliability of the statistical analysis, this study conducted

general assumption tests for mean analysis in addition to the test of three

unique-to-ANCOVA assumptions.

Conceptually, the assumption of independence is valid in the present study, since

each company registered with ISO-14001 EMS individually and voluntarily. After log

transformation (base e), the assumptions of equal variance and normality were qualified

for the three dependent variables in the study period (p > 0.0167). In addition, this study

has examined the heteroscedasticity (residuals of dependent variable verses predicted),

multicollinearity (VIF < 10), and standardised residuals test for outliers (less than ±3). No

violations were identified.

The three unique-to-ANCOVA assumptions were tested. The assumption of linearity

was qualified (Pearson Correlation test: ‘r’ > ±0.20). The assumption of independence

was valid in the present study because the covariate data were collected before the

treatments (ISO-14001 registration). The assumption of equal regression slopes was valid

for the ROR and ROA (F < F*), but violated (F > F*) for OPR. Therefore, MANOVA

and ANOVA were invited for the analysis of OPR for Hypothesis 3. As Huck (2004,

p.405) suggested that in the case when the equal-slope assumption is rejected, “there are

several options open to the researcher … data can be transformed … Or, the researcher

can decide to pay no attention to the covariate data and simply use an ANOVA to

compare groups on the dependent variable”. The main purpose of inviting MANOVA

was to make final conclusion based upon the higher statistical power.

4 Results

4.1 Distribution of ISO registration

The distribution of ISO companies before and after sampling revealed similar patterns by

industries, by years, and by the size of company.

4.1.1 Distribution of ISO registration by industries

Thirty-four out of 44 industries have companies registered with ISO-14001. Of the 1279

S&P companies, 214 (16.73%) obtained one or more ISO certifications through August,

2005. Eighty-one of these companies were selected for data analysis (see Table 4).

Of the 34 industries, heavy equipment and trucks, auto, chemical industries ware

ranked as the top three industries with the highest registration both before and after

sampling (see Table 4). Thus the sampling process adequately captured the industrialconcentration of ISO-14001 registration. It might suggest that industries with higher

consumption of natural resources are more environmentally sensitive than industries with

lower consumption of natural resources.




Table 4 Distribution of ISO companies by industries

Industry1279 S&Pcompanies

214 ISOcompanies

Percentage of

all ISOcompanies

(%)

81 sampled ISO

companies

Percentageof sample

(%)

Advertising 27 0 0.00 0 0.00

Aerospace and defense 27 9 4.21 3 3.70

Agribusiness industry 20 1 0.47 0 0.00

Airlines industry 21 0 0.00 0 0.00

Alcoholic beverages and tobacco 16 0 0.00 0 0.00

Apparel and footwear industry 31 1 0.47 1 1.23

Autos and auto parts 29 19 8.88 9 11.11

Banking industry 29 1 0.47 1 1.23

Biotechnology industry 13 1 0.47 1 1.23

Broadcasting and cable industry 31 0 0.00 0 0.00Chemicals: basic industry 62 18 8.41 6 7.41

Communications equipment 32 5 2.34 3 3.70

Computers: hardware 30 9 4.21 3 3.70

Computers: storage and peripherals 24 5 2.34 2 2.47

Computers: software industry 49 1 0.47 1 1.23

Electric utilities industry 33 6 2.80 1 1.23

Environmental management 10 2 0.93 1 1.23

Financial services: diversified 27 0 0.00 0 0.00

Foods and beverages 56 7 3.27 3 3.70

Healthcare: facilities industry 15 1 0.47 0 0.00

Healthcare: managed care 15 0 0.00 0 0.00

Healthcare: products and supplies 45 6 2.80 2 2.47

Heavy equipment and trucks 62 21 9.81 3 3.70

Homebuilding industry 21 3 1.40 0 0.00

Household durables industry 18 5 2.34 2 2.47

Household nondurable industry 25 3 1.40 3 3.70

Industrial machinery industry 39 15 7.01 8 9.88

Insurance: life and health 18 0 0.00 0 0.00

Insurance: property-casualty 35 0 0.00 0 0.00

Investment services industry 16 0 0.00 0 0.00

Metals: industrial industry 22 8 3.74 4 4.94

Movies and home entertainment 22 3 1.40 1 1.23

Natural gas distribution 34 6 2.80 1 1.23

Oil and gas equipment 24 5 2.34 1 1.23

Oil and gas: production and 45 8 3.74 4 4.94

Paper and forest products 28 10 4.67 4 4.94

Healthcare: pharmaceuticals 27 5 2.34 3 3.70Publishing industry 25 3 1.40 1 1.23

Semiconductor equipment 26 10 4.67 2 2.47

Semiconductors industry 34 14 6.54 5 6.17

Supermarkets and drugstores 32 0 0.00 0 0.00

Tele-wireless industry 21 1 0.47 1 1.23

Tele-wire line industry 21 1 0.47 1 1.23

Transportation: commercial 42 1 0.47 0 0.00

Total percentage 1279 214 100.00 81 100.00



222 J. Zhao

4.1.2 Distribution of ISO registrations by years (from 1996 to 2002)

Except for the year 1996 in which 73 registrations were recorded, the remaining six years

recorded fewer than 24 registrations per year (see Figure 3). The sampled 81 ISO

companies demonstrated a similar time-related pattern of distribution as the 214 ISO

companies for the period from 1996 to 2002. This time-related pattern indicated a

decrease of registration after 1996 (see Figure 3). The logical explanation might be that

ISO registration is a voluntary process, rather than mandatory, and companies did not

really see the expected effects of adoption and implementation of ISO-14001 after

initiated in 1996. Therefore, companies might choose other types of management system

to improve their organisations.

Figure 3 The distribution of sampled ISO-14001 registration by years

4.1.3 Distribution of ISO registrations by size of company

Of the 81 ISO companies, 39 (48.15%) belong to the S&P large cap, 27 (33.33%) to the

mid cap, and 15 (18.52%) to the small cap (see Table 5). This distribution of ISO

registration versus company size may suggest that larger companies are more likely to

register with ISO than smaller companies (King and Lenox, 2002). Early adopters

represent over fifty percent of large companies as well. Larger companies may

be more capable of investing in management systems, because larger companies may

have more facilities over which to distribute the cost of investment (King and Lenox,

2002; Russo, 2007).

Table 5 Size and early adoption

T ± 2 period T ± 3 period

Size cap Number of company Early Number of company Early

Large 39 24 22 14

Mid 27 15 16 9

Small 15 6 7 4

Total 81 45 45 27




4.2 Results from ANCOVA analysis

4.2.1 Results of ROR (H1) and ROA (H2)

Using the averaged pre-scores of ROR and ROA as covariate variables to control for

initial differences, and post-scores of ROR and ROA as dependent variables, ISO

registration as the independent variable (binary variable: ISO company as ‘1’, otherwise

‘0’) at significant level α = 0.0167, holding other factors constant, on average, the

difference of adjusted means between the ISO companies and the matched comparison

companies for ROR was significantly different in the (T ± 2) period (p = 0.012), but not

significantly different in the (T ± 3) period (p = 0.121). For ROA, the differences of the

adjusted means were not significant for both periods (p = 0.041, p = 0.194). Furthermore,

the adjusted means for ROR and ROA were lower in the ISO companies than the

matched comparison companies, indicating that there was a reduction in ROR and ROA

for the ISO companies compared to the matched comparison companies. Therefore, H1was negatively supported for the (T ± 2) period, but not supported for the (T ± 3) period.

H2 was not supported for either period (see Table 6).

Table 6 Effect of adjusted means using ANCOVA

T ± 2 (N = 162) T ± 3 (N = 90) Data means logtransformation(Base e) ISO Match P-value Power ISO Match P-value Power

ROR

Covariate (ROR-Pre) 1.805 1.791 0.000* 1.000 1.802 1.720 0.000* 0.998

Covariate (ROA-Pre) 0.389 0.063 0.553 0.037

Posttest 1.601 1.846 1.667 1.825

Adjusted 1.594 1.852 0.012* 0.548 1.637 1.855 0.121 0.197

ROA

Covariate (ROA-Pre) 1.873 1.859 0.000* 1.000 1.910 1.827 0.000* 1.000

Posttest 1.624 1.833 1.705 1.837

Adjusted 1.619 1.838 0.041 0.362 1.679 1.864 0.194 0.135

Size effects (Size*ISO)

Covariate (ROR-Pre) 0.000* 1.000 0.000* 0.995


Interaction (ROR) 0.234 0.174 0.037 0.456

Adjusted 1.539 1.865 0.003* 0.716 1.492 1.886 0.010* 0.572

Covariate (ROA-Pre) 0.000* 1.000 0.000* 1.000

Interaction (ROA) 0.167 0.223 0.041 0.437

Adjusted 1.550 1.849 0.010* 0.579 1.535 1.896 0.021 0.469

Early effects (Early*ISO)

Covariate (ROR-Pre) 0.000* 1.000 0.000* 0.999


Interaction (ROR) 0.082 0.255 0.199 0.131

Adjusted 1.575 1.853 0.007* 0.629 1.599 1.854 0.073 0.271

Covariate (ROA-Pre) 0.000* 1.000 0.000* 1.000

Interaction (ROA) 0.041 0.362 0.200 0.131

Adjusted 1.593 1.836 0.021 0.465 1.629 1.852 0.117 0.201

Note: * Statistically significant at α Level: α = 0.0167.



224 J. Zhao

4.2.2 Results of H4 for ROR and ROA

Using the averaged pre-scores of ROR and ROA as the covariate variables, early

adoption (binary variable: early adopter as ‘1’, otherwise ‘0’) together with ISO

registration as two independent variables (α = 0.0167), and holding other factors

constant, on average, the results showed that the interaction of early adoption and ISO

registration is not significantly related to financial performance (ROR: p = 0.082 in T ± 2

period, p = 0.199 in T ± 3 period; ROA: p = 0.041 in T ± 2 period, p = 0.20 in T ± 3

period). Therefore, H4 for ROR and ROA was not supported (see Table 6).


As noted above, this study used the actual number of years of ISO registration as the

independent variable, and post-scores of ROR and ROA for the year 2003 were used as

the new dependent variables. The data for H5 test was re-sampled and finalised as 58pairs of ISO companies and their matched comparison companies ( N = 116). Holding

other factors constant (α = 0.0167), on average, the results showed that the mean

difference of the interaction between number of years of registration and ISO registration

was not significantly related to financial performance (for ROR: p = 0.423; for ROA:

p = 0.279). Therefore, H5 for ROR and ROA was not supported (see Table 7).

Table 7 ANCOVA test on length (H5) (length: Y = 1996 – 2003 N = 116)

Return on Revenue (ROR) Return on Assets (ROA) Data

(fiscal yr.)

Sample

size ISO Match P-value Power ISO Match P-value Power

Y-1 (02/03) N1 = 6 1.902 1.815 1.852 1.914

Y-2 (01/02) N2 = 22 1.034 1.912 0.908 1.864

Y-3 (00/01) N3 = 12 1.246 1.685 0.852 1.796

Y-4 (99/00) N4 = 18 1.496 2.227 1.600 2.059

Y-5 (98/99) N5 = 8 1.625 2.029 1.475 2.022

Y-6 (97/98) N6 = 6 1.635 1.756 1.641 1.827

Y-7 (96/97) N7 = 22 1.690 1.657 1.646 1.664

Y-8 (1996) N8 = 22 1.874 1.750 1.955 1.647

Adjusted N = 116 1.563 1.854

Standard deviation

ISO = 0.298

Match = 0.194

1.491 1.849

Standard deviation

ISO = 0.406

Match = 0.150

ROR Pre 0.000* 0.994Covariance

ROA Pre 0.611 0.031

ROA Pre 0.000* 0.966

Interaction (ISO*Years) 0.423 0.255 Interaction 0.279 0.339

Note: * Statistically significant at α Level: α = 0.0167.


Using the pre-scores of ROR and ROA as covariate variables, ISO registration and

company size were used as two independent variables (S&P 600 small cap as ‘1’,

S&P 400 mid cap as ‘2’, and S&P 500 large cap as ‘3’). Holding other factors constant

(α = 0.0167), on average, the results showed that the means of interaction between size

and ISO registration was not significantly related to financial performance (for ROR:




N = 162, p = 0.234 in T ± 2 period, and N = 90, p = 0.037 in T ± 3 period; for ROA:

N = 162, p = 0.167 in T ± 2 period, and N = 90, p = 0.041 in T ± 3 period). Therefore, H6

for ROR and ROA was not supported (see Table 6).

4.3 MANOVA and ANOVA analyses

This study used the change scores (between pre-score and post-score) of ROR, ROA, and

OPR as the three dependent variables, ISO registration as the independent variable, and

the same control as ANCOVA analyses for early adoption, length of registration, and

company size. Holding other factors constant (α = 0.05), on average, the results of

MANOVA and ANOVA show that H1 and H2 were negatively supported in the (T ± 2)

period, but not supported in the (T ± 3) period. The results of H3, H4, H5 and H6 were

consistent with ANCOVA (see Tables 6, 7 and 8).

Compared to the result from ANCOVA analysis, in which H2 was not supported inthe (T ± 2) period, the result from MANOVA and ANOVA provided higher statistical

power (0.714 > 0.362, see Table 6), therefore, this study accepted the result from

MANOVA and ANOVA that H2 was negatively supported in the (T ± 2) period.

Table 8 Verification using MANOVA test

T ± 2 α = 0.05 N = 162 T ± 3 α = 0.05 N = 90

DV F P Decision F P Decision

H1 ROR 6.44 0.0121* Reject 2.82 0.0968 Accept

H2 ROA 4.15 0.0433* Reject 2.04 0.1564 Accept

H3 OPR 0.81 0.3689 Accept 0.47 0.4953 Accept

ROR 2.89 0.0909 Accept 1.84 0.1785 Accept

ROA 3.80 0.0530 Accept 1.53 0.2195 Accept

H4

OPR 0.10 0.7547 Accept 0.07 0.7940 Accept

ROR 1.60 0.2057 Accept 3.16 0.0476 Reject

ROA 1.77 0.1735 Accept 2.41 0.0959 Accept

H6

OPR 1.30 0.2758 Accept 0.83 0.4378 Accept

Path dependent learning test N = 116

(Hypothesis 5: Effect of length of registration)

DV F-value P-value Decision

ROR 0.96 0.4665 Accept

ROA 0.84 0.5603 Accept

OPR 0.45 0.8678 Accept

Note: * p < 0.05 significant level.

5 Conclusion

The quasi-experimental design with strict matching protocol provides better control for

alternative explanations than previous studies on ISO-14001. It is confident to conclude

that the difference of financial performance between ISO companies and matched



226 J. Zhao

comparison companies is likely attributable to the adoption and implementation of

ISO-14001. By employing the ANCOVA and testing the baseline value of both ISO andnon-ISO companies, this study gained better control for international validity.

The results2 of statistical analysis can be summarised as follows: ISO-14001

registration was associated with decreases in ROR and ROA (H1 and H2) for the (T ± 2)

period, but not supported for the (T ± 3) period. OPR (H3) was not supported. Early

adoption, length of registration and size of company (H4, H5 and H6) were not related to

the three dependent variables.

The data on ISO companies collected in this study revealed certain patterns. First,

there appears to be a relationship between ISO registration and industry groups,

indicating a heavy concentration of registration in environmentally-sensitive industries.

Second, larger companies (S&P LargeCap-S&P 500) appear to be more likely to register

with ISO-14001 earlier than smaller companies (S&P MidCap-S&P 400, and S&P

SmallCap-S&P 600).

5.1 Limitations and recommendations

5.1.1 Why do ROR and ROA significantly decrease in the (T + 2) period?

The interpretation of the reversed impact on ROR and ROA in the (T ± 2) period is

two-fold: First, costs of adoption and implementation are not measured in this study; such

costs would have been accounted as expenses in business operations and would therefore

reduce net OPR. No published articles were found that systematically collected

accounting data on the costs of adoption and implementation. Subjective and self-claimed

estimates of setup costs were reported (Hamschmidt and Dyllick, 2001; NDEMS, 2003).

These authors noted that their numbers were estimated figures, not accounting data.

Future research may want to examine the cost related issues in the study of ISO-14001.

Second, the quality of adoption and implementation was not examined in the present

study; such quality should have been used as the measure of the third element of dynamic

capability theory – ability of integration. This might be the reason to explain why are

Early Adoption, Length of Registration and Company Size not significantly related to

financial performance. Future study should examine the dynamic capability theory in a

holistic way. Or, the theory itself needs to be redefined in order to explain the effect

of ISO-14001.

5.1.2 Why are differences not significant in the (T + 3) period for RORand ROA?

As noted earlier, the sample size for the (T ± 3) period was considerably lower than for

the (T ± 2) period, consequently, reduced statistical power is the most likely explanationfor the lack of statistical significance (see Table 8).

Another possible explanation is that the (T ± 2) period might be a transitional period

of organisational change resulting from organisational adoption and adaptation of a new

management system (ISO-14001 EMS). The non-significant differences in the (T ± 3)

period may indicate that the costs of adoption and implementation become less of a

burden after the transitional period. Therefore, the possible speculation is that the

decreased ROR and ROA in (T ± 2) period may ‘catch-up’ in the (T ± 3) period.




Further investigation regarding a two-year period of turnaround is an area for

future research. A two-year start-up period is comparable to 2.2 years presented by

Hamschmidt and Dyllick (2001), whose research was a self-report survey on payback of

ISO-14001 in Switzerland (N = 158). This speculation of two-year turnaround is

consistent with Corbett et al.’s (2005) findings regarding payback following the

implementation of ISO-9000.

5.1.3 Why are the results for OPR not significant in both periods?

One explanation might be related to a ‘lagged effect’ of ‘pays to be green’. As mentioned

earlier, this study is based on a theoretical assumption: ‘Green Must Pay’. However,

when does it pay is still unknown. Future research may want to identify and examine

some controlling drivers that can improve the quality of implementation, and speed up

the process from ‘Costs to Be Green’ to ‘Pays to Be Green’.

5.1.4 Issues regarding research design

Despite the fact that this study is unique in quasi-experimental design in the study

of ISO-14001, comparing to the true experimental design, the weakness of

quasi-experimental design needs to be addressed, even when comparison groups are

available (Campbell and Stanley, 1963). Two issues ring the cautious bell for future

research: the first issue is on construct validity. The possession of ISO certification has

limited construct validity as a measure of EMS. Since data on implementation has not

been collected and measured in this study, and the binary nature of ISO registration

(registered ‘1’, non registered ‘0’) may not be a valid measure capable of representing the

status of EMS. Following this line of reasoning, ISO certification has limited construct

validity for differentiating whether companies are practicing environmental management.The second issue is related to the external validity. Although, the present study

extended the scope of industries used in the previous research, the data in this study is

still limited to S&P capitalisations within USA. Companies selected are relatively large

compared to most of the companies that are not listed by S&P NetAdvantage database.

Therefore, the scope of the generalisability of this study is limited.

Future research may want to replicate this study, but comparing S&P companies to

non-S&P companies within USA, to see if similar results hold true. In addition, since

ISO-14001 has been successfully diffused globally, future research may want to replicate

the current research design, and extend the research scope from a single country to

multiple countries with similar economic levels but different social, political, cultural,

and natural environments.

5.1.5 Methodological issues

Despite quasi-experimental design using the matched control group to control for

the initial differences of dependent variables, several issues still represent potential

threats to the validity of the present study, such as diluted effect, lagged effect, and

overlapped effect.

Diluted effect occurs when a large company has many facilities, but only a small

portion of those facilities registered with ISO-14001, then the detected effect is diluted.

The present study originally planned to collect the proportional data (number of



228 J. Zhao

registered facilities divided by total number of facilities). However, due to the limitations

of data source (i.e., increase or decrease of total number of facilities verses number of registered facilities may vary from year to year), therefore, this study could not obtain the

proportional data to control for the dilution. Future research may control for this dilution

by collecting the proportional data.

Lagged effect in this study may be the cause of the decreased financial performance

for ISO companies in the (T ± 2) period resulting from the inability of organisations to

absorb, assimilate, and integrate the full function and value of a new management system

such as ISO-14001 into their existing management system. As mentioned earlier, ability

of integration may be considered the key element of dynamic capability theory to explain

the performance of ISO-14001.

Overlapped effect occurs when both the ISO companies and the matched comparison

companies may have installed other environmental programme(s) prior to ISO-14001

registration (i.e., Ford and GM). Therefore, the findings of this study may reflect acombined (overlapped) effect from both existing EMS and ISO-14001 EMS. Future

research may need to have two groups of matched comparison companies in order to

control for this threat: one group uses other type of EMS, and one group does not use

EMS; then compare the experimental group (ISO group) with each of these two groups.

Acknowledgement

This paper is a modified version based upon my PhD dissertation under the supervision

of Dr. Dennis Heaton, Dr. Maxwell Rainforth, Dr. Jane and Dr. Robert Roney. I would

like to thank them for their support, suggestions and advice.

I am also deeply appreciative to the following people for substantive advice and

suggestions they gave me: Dr. Michael V. Russo (Professor at Lundquist College of

Business, University of Oregon, OR), Dr. Andrew A. King (Associate Professor at Tuck

School of Business at Dartmouth, NH), Dr. Michael J. Lenox (Associate Professor at

Fuqua School of Business, Duke University, Durham, NC), Dr. Richard Andrews

(Associate Professor at University of North Carolina at Chapel Hill, NC), and Dr. Petra

Christmann (Assistant Professor at Rutgers Business School, Rutgers University, NJ).

The author gratefully acknowledge the helpful and constructive comments of

anonymous reviewers which helped to improve the presentation of the paper

considerably.

References

Bansal, P. and Bogner, W.C. (2002) ‘Deciding on ISO 14001: economics, institutions, andcontext’, Long Range Planning, Vol. 35, No. 3, pp.269–290.

Brigham, E.F. (1995) Fundamentals of Financial Management , Orlando: The Dryden Press.

Campbell, D.T. and Stanley, J.C. (1963) Experimental and Quasi-Experimental Designs for Research, Boston: Houghton Mifflin Company.

Christmann, P. and Taylor, G. (2003) ‘Globalization and the environment: the role of ISO 14000’,Paper Presented at the 2000 Academy of Management Annual Conference , Toronto, Canada.

Corbett, C.J., Montes-Sancho, M.J. and Kirsch, D.A. (2005) ‘The financial impact of ISO 9000certification in the United States: an empirical analysis’, Management Science, Vol. 51, No. 7,pp.1046–1059.




Corporate Risk Management (2006) ‘Worldwide numbers of ISO 14001’,

http://www.ecology.or.jp/isoworld/english/analy14k.htm (23 May 2006).

Delmas, M.A. (2006) ‘Barriers and incentives to the adoption of ISO 14001 by firms in the UnitedStates’, https://www.law.duke.edu/journals/delpf/articles/ (2 April 2006).

Edwards, A.J. (2004) ISO 14001 Environmental Certification Step by Step, Boston: ElsevierButterworth Heinemann.

Egri, P.C. and Herman, S. (2000) ‘Leadership in the North American environmental sector: values,leadership styles, and contexts of environmental leaders and their organizations’, Academy of Management Journal, Vol. 43, No. 4, pp.571–605.

Geffen, A.C. and Rothenberg, S. (2000) ‘Suppliers and environmental innovation: the automotivepaint process’, International Journal of Operations & Production Management , Vol. 20,No. 2, pp.166–186.

Hamschmidt, J. and Dyllick, T. (2001) ‘ISO 14001: profitable – yes! But is it eco-effective?’,Greener Management International (GMI), Vol. 34, pp.43–54.

Huck, S.W. (2004) Reading Statistics and Research, 4th ed., Boston: Pearson Education Inc.

Institute of Environmental Management & Assessment (IEMA) (2005) ‘Revision to ISO 14001:1996 – what the publication of the 2004 edition means for organizations using the standard’,http://www.iema.net/xar/library/ISO14001_changes/ (30 June 2005).

International Organization for Standardization (ISO) (2002) Environmental Management: The ISO14000 Family of International Standards, Geneva, Switzerland: ISO.

International Organization for Standardization (ISO) (2004) ‘ISO 14001 2004’,http://www.iso.org/iso/en/ISOOnline (9 July 2005).

Jiang, R.J. and Bansal, P. (2003) ‘Seeing the need for ISO 14001’, Journal of Management Studies,Vol. 40, No. 4, pp.1047–1067.

King, A.A. and Lenox, M.J. (2002) ‘Exploring the locus of profitable pollution reduction’, Management Science, Vol. 48, No. 2, pp.289–299.

King, A.A., Lenox, M.J. and Terlaak, A. (2005) ‘The strategic use of decentralized institutions:

exploring certification with the ISO 14001 management standards’, Academy of Management Journal, Vol. 48, No. 6, pp.1091–1106.

Melnyk, S.A., Sroufe, R.P., Calantone, R.L. and Montabon, F.L. (2002) ‘Assessing theeffectiveness of US voluntary environmental programmes: an empirical study’, International Journal of Production Research, Vol. 40, pp.1853–1878.

NDEMS Final Report (2003) ‘The effects of ISO 14001 and other environmental managementsystems’, http://ndems.cas.unc.edu/final_report.htm (18 December 2006).

Parker, L.D. (2000) ‘Environmental costing: a path to implementation’, Australian Accounting Review, Vol. 10, No. 3, pp.43–51.

Russo, M.V. (2002) ‘Does ISO 14001 reduce toxic emissions? Early evidence from the electronicsindustry’, Paper Presented at the National Pollution Prevention Roundtable, Portland,Oregon, April.

Russo, M.V. (2007) ‘Explaining the impact of voluntary management standards: a dynamic

capabilities perspective on process and learning’, Business Strategy and the Environment , Forthcoming.

Russo, M.V. and Harrison, N.S. (2005) ‘Organizational design and environmental performance:clues from the electronics industry’, Academy of Management Journal, Vol. 48, No. 4,pp.582–593.

Standard & Poor’s NetAdvantage Financial Database Room (2005) http://www.library.hbs.edu/ financialdatabaseroom.html (accessed in 2006).

Starik, M. and Marcus, A.A. (2000) ‘Introduction to the special research forum on the managementof organizations in the natural environment: a field emerging from multiple paths, with manychallenges ahead’, Academy of Management Journal, Vol. 43, No. 4, pp.539–547.



230 J. Zhao

Stead, E.W., Stead, J.G. and Starik, M. (2004) Sustainable Strategic Management , New York: M.E.

Sharpe Inc.Stevens, J.P. (1997) Intermediate Statistics: A Modern Approach, 2nd ed., London: Lawrence

Erlbaum Associates, Publishers.

Teece, D., Pisano, G. and Shuen, A. (1997) ‘Dynamic capabilities and strategic management’,Strategic Management Journal, Vol. 18, pp.509–533.

Toffel, M.W. (2003) ‘The growing strategic importance of end-of-life product management’,California Management Review, Vol. 45, No. 3, pp.102–130.

Von Zharen, W.M. (2001) ISO 14001: Positioning Your Organization for Environmental Success,Rockville, Maryland, Government Institutes Press.

Wilson, M. and Lombardi, R. (2001) Globalization and Its Discontents: The Arrival of Triple-Bottom-Line Reporting, Toronto: Ivey Business Journal.

Notes

1 The Restricted Group was also selected by defining matching portfolio for assets’ size(50%–200%) used in Corbett et al.’s (2005) study. This group is finalised as 63ISO companies matched with 63 non-ISO companies (N = 126) with three dependentvariables available.

2 The present study also conducted statistical analyses for the original group and restrictedgroup. The results from the original group are the same as main group in both (T ± 2) and(T ± 3) periods. Between the main group and restricted group, the result is the same in (T ± 3)period, but different in (T ± 2) period on ROA ( p = 0.0892).

ISO 14001 and Financial Performance (2008) 21pp

Documents