W
.

flay

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o:
. .L

It». 1...

”.1 r 1X
.an...

. baby“,

Vtfim.

. if... .3. hr.

Evita-.1
:1.“ 3
.1323. .3.
kn: .1... w.

. v.11..v...n.!$..
.. l:l..n5........2.
\l‘...
. .31

)JVn.’x .\ .
:55ra... .

53:...
..».....t..n...h«v

 

..
.m Wu....:.=.2.. ”$3.3m? 3%
.

% . at...
fin... .n... 5%“ {$6qu
”.naflhmwn. afin £93..th

n :2 til;

3.
L": .;!g

h

4
t5....
.1. .flfinm.

fight)”.

.t.3l$‘:'t ‘33.}! Tlgtig‘u
lit......s 1.3.1.525. “Halt. {tins}.
1.13.52 H.102...
iv...xei. .Ktntht It .4."

.fin. 3.5!“. MINNIE. .Ittttcnr. :5
hr. ziltt. . .
inlwfltrlrnnflnuufl In 1...... ISO-i
. . x L :N.ltn.b.!hfl
air.Ԥt Iufldrluvlir.la:w
: .1..t.ntn.1.. . a}. a“... “an”:
r. In 5.

1 a
. ’{Inl‘l 91...?! I!
.t :52. .....II.:......

.115“

fi...:....:
s 1.6: fir. 1135:. n52!
1:10;..3.‘ >3

 

This is to certify that the
dissertation entitled

ANALYSIS OF COUNTERFEIT RISKS AND
DEVELOPMENT OF A
COUNTERFEIT PRODUCT RISK MODEL

presented by

JOHN WILLIAMS SPINK

has been accepted towards fulfillment
of the requirements for the

 

 

 

 

 

Doctoral degree in Packaging
// "W -
7 Major Professor’s Signature
é/fla/a 7
I I '
Date

MSU is an Afiirmative Action/Equal Opportunity Employer

ANALYSIS OF COUNTERFEIT RISKS AND DEVELOPMENT OF A
COUNTERFEIT PRODUCT RISK MODEL

By

John Williams Spink

A DISSERTATION

Submitted to
Michigan State University
in partial fiilfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Packaging

2009

ABSTRACT

ANALYSIS OF COUNTERFEIT RISKS AND DEVELOPMENT OF A
COUNTERFEIT PRODUCT RISK MODEL

By
John Williams Spink

Product counterfeiting is growing in both scope and scale. There is a need to take
a holistic, all-encompassing approach to the anti-counterfeit strategy, including the
development of a Counterfeit Product Risk Model. This research process collaborated
and leveraged a wide variety of academic and industry expertise utilizing a literature
review and interdisciplinary peer consultation to develop the Counterfeit Product Risk
Model for consumer products. The range of disciplines for the research included:
Packaging, Food Safety, Criminal Justice, Supply Chain/Logistics, Risk Analysis/Risk
Assessment, Food Law, Food Safety, Food Defense, Intellectual Property Rights Law,
Political Science, and Social Science. For example, the Criminal Justice concepts include
“the chemistry of a crime” and “the crime triangle.” The Counterfeit Product Risk Model
focuses on the probability portion of a traditional probability versus severity matrix, uses
qualitative ranking, and due to the nature of the risk and the data, emphasizes extensive
use of expert panels.

This research defines five factors related to counterfeiting: Counterfeit-History,
Counterfeit-Attractiveness, Counterfeit-Ability, Counterfeit-Hurdles, and Market Profile.
The Model defines the derivation and integration of sub-factors, which “roll-up” to

determine the rank of the factors.

The model was then validated using a survey of 33 industry and agency experts.
The survey included 17 ratings by people at a Corporate or Vice-Presidential level, and
included six $1-$5 billion revenue companies and sixteen over $5 billion revenue
companies. A broad and representative balance of industries was included: food,
beverage, healthcare, pharmaceutical, medical device, law, finance, insurance, risk,
consumer electronics, software, industrial original equipment manufacturers, and
consumer packaged goods.

“Assessing Agreement” analysis was conducted on the surveys and the
interpretation of the result was an “almost perfect agreement” with the model. F leiss’
Kappa analysis was conducted to assess agreement over random chances, and this result
was also an “almost perfect agreement.” The research included a Case Study to
demonstrate the use of the Model.

This research provides a valuable analysis of anti-counterfeit strategy, including
an extensive look into the historical information. It provides a theoretically supported
Counterfeit Product Risk Model that will assist in disrupting the “chemistry of the

9,

crime.

Copyright by
JOHN WILLIAMS SPINK
2009

DEDICATION

From my 1991 MSU Packaging Master of Science Dedication:
This thesis is dedicated to my parents, sister,
grandmothers and great aunts who encouraged and
supported my graduate education.
They instilled in me a desire to learn and raise
myself to a higher level. They instilled in me the desire to
fulfill my high self-expectations. They instill in me the
idea that I can do anything.
This dissertation is dedicated to my parents who look to make the world a better
place. This dissertation is dedicated to my children, Celia and Kirill, who drive me to
want to make the world a better place. But mostly, this is dedicated to my lovely wife

who was my constant companion, supporter, and co-innovator. She helped refine and

support the shared vision that is evidenced in this document and my life’s work.

ACKNOWLEDGEMENTS

I would like to thank the following people, without whom this work would not
have been possible:

Dr. Bruce Harte, for his guidance, expertise, support, camaraderie, and endless
patience while serving as Advisor. I am indebted to him for first, helping identify this as
a research topic, and second, for both giving me the free reign to explore the topic but
also to drive me towards applicable ends. This work, and the expanding research area of
Anti-Counterfeit Strategy, is a tribute to his vision.

Dr. Ed Mather, for his unyielding support and encouragement of all my
endeavors. I am especially indebted to him for originally bringing me back to MSU in a
role with the National Food Safety and Toxicology Center, and specifically with the
opportunity to develop and teach courses in the On-Line ProMS Food Safety Program.
He has been instrumental through encouragement and support of Food Fraud and Food
Counterfeiting as a unique scientific concept.

Dr. Robb Clarke and Dr. Laura Bix both provided expert insight fi'om their areas
of research, and both were instrumental and tireless in their encouragement and guidance
of my development as an academic and as a scholarly researcher.

Four other MSU Faculty Members were especially import to my ongoing
research, and they are Dr. Julie Funk of Food Safety, Dr. Scott.Winterstein of the
National Food Safety and Toxicology Center, Neal Fortin of the International Food Law

Institute, and Dr. Ed McGarrell of the School of Criminal Justice.

vi

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................... viii
LIST OF FIGURES ......................................................................................................... x
INTRODUCTION .............................................................................................................. 1
1.0 LITERATURE REVIEW ....................................................................................... 4
1.1 Counterfeit Definition Overview ........................................................................ 5
1.2 Counterfeit Definition Detail .............................................................................. 6
1.3 Scale of the Threat ............................................................................................ 12
1.4 Scope of the Threat ........................................................................................... 16
1.5 Counterfeit Drugs .............................................................................................. 17
1.6 Countermeasure Action and Deterrence ........................................................... 21
1.7 Summary ........................................................................................................... 28
2.0 RESEARCH ANALYSIS ..................................................................................... 29
2.1 Risk Factor Selection Analysis ......................................................................... 30
2.2 Risk Concept Research ..................................................................................... 35
2.3 Risk Model Research ........................................................................................ 60
3.0 MODEL DEVELOPMENT .................................................................................. 97
3.1 Model Overview ............................................................................................... 97
3.2 Research Introduction ....................................................................................... 99
3.3 Propositions ..................................................................................................... 100
3.4 Model Development Summary ....................................................................... 112
4.0 EXPERT SURVEY ............................................................................................ '113
4.1 Methods — Expert Survey ................................................................................ 113
4.2 Expert Survey Results and Discussion ...................................................... 119
5.0 CASE STUDY .................................................................................................... 132
6.0 CONCLUSION ................................................................................................... 138
6.1 For Further Research ....................................................................................... 140
APPENDICES ................................................................................................................ 142
APPENDIX A -— Results of Expert Survey ................................................................. 143
APPENDIX B — Raw Statistical Analysis .................................................................. 144
APPENDIX C -— Survey Responses — Raw Comments ............................................... 146
APPENDIX D — Case Study Roll-Up Detail .............................................................. 150
APPENDIX E — Case Study Example of Automated Function .................................. 159
REFERENCES ............................................................................................................... 162

vii

LIST OF TABLES

Table 1. Estimated Value of Counterfeit Goods Based on Success Rate of Seizures ...... 15

Table 2. List of Counterfeit Products from GBLAAC ..................................................... 16
Table 3. List of Counterfeit Products from WCO ............................................................. 16
Table 4. A General Probability versus Severity Matrix .................................................... 51
Table 5. Carver+Shock Software Tool: Shock Scale Descriptions .................................. 70
Table 6. Zurich Hazard Analysis Matrix of Severity versus Probability .......................... 74
Table 7. Suggested Mishap Probability Levels for Specific Events and Occurrences in

Mil Std 882D ..................................................................................................................... 75
Table 8. Suggested Mishap Severity Categories in Mil Std 882D ................................... 76
Table 9. Example of mishap risk assessment values in Mil Std 882D ............................. 77
Table 10. Example mishap risk categories and mishap risk acceptance levels ................ 77
Table 11. Fire Danger Rating and Color Codes from the WP AS System ........................ 89
Table 12. Interpretation of Agreement for Expert Panel Survey Results ....................... 117
Table 13. Subject Classification — Title of the Expert Survey Subjects ......................... 120
Table 14. Subject Classification — Industry of the Expert Survey Subjects ................... 121

Table 15. Survey Classification — Company Revenue of the Expert Survey Subjects... 122

Table 16. Results of Survey Question 2 — Subject Expertise (Self-Assessment) for the

Expert Survey Subjects ................................................................................................... 122
Table 17. Survey Results of the Expert Survey Subjects ................................................ 124
Table 18. Case Study Overall Counterfeit Risk for CFSI ............................................... 137
Table 19. Raw Statistical Analysis-Inputs and Overall Agreement ......................... 144
Table 20. Raw Statistical Analysis-Inputs and Overall Agreement .............................. 145
Table 21. Case Study Overall Counterfeit Product Risk, Summary ............................... 150

viii

Table 22
Table 23
Table 24
Table 25
Table 26
Table 27
Table 28
Table 29
Table 30
Table 31
Table 32
Table 33
Table 34
Table 35
Table 36
Table 37
Table 38
Table 39
Table 40
Table 41
Table 42
Table 43

. Case Study Factor 1 Counterfeit-History, Summary ...................................... 151
. Case Study Factor 2 Counterfeit—Ability, Summary ....................................... 151
. Case Study Factor 3 Counterfeit-Attractiveness, Summary ........................... 152
. Case Study Factor 4 Counterfeit-Hurdles, Summary ..................................... 152
. Case Study Factor 5 Market Profile, Summary .............................................. 153
. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.1 ............... 153
. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.2 ............... 154
. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.3 ............... 154
. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.1 ............... 154
. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.2 ............... 155
. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.3 ............... 155

. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.1 155

. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.2.... 156
. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.3 156
. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.1 .............. 156
. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.2 .............. 157
. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.3 .............. 157
. Case Study Factor Derivation, Market Profile, Sub-factor 5.1 ....................... 157
. Case Study Factor Derivation, Market Profile, Sub-factor 5.2 ....................... 158
. Case Study Factor Derivation, Market Profile, Sub-factor 5.3 ....................... 158
. Case Study Example of Automated Total Risk Determination ...................... 160
. Model Development — Example of Automated Summary of Total Rank ...... 161

LIST OF FIGURES

Figure 1. Development of the Risk Model Using a Multiple Component Research

Approach ............................................................................................................................. 3
Figure 2. The Risk Analysis Concept and its Components .............................................. 37
Figure 3. Counterfeit Product Risk Model Overview ....................................................... 98
Figure 4. Example of Probability versus Severity Matrix .............................................. 101

Figure 5. Example of Factors Defining Probability for the Overall Counterfeit Risk.... 108

Figure 6. Example of the Sub-F actor Derivation of the Market Profile Factor .............. 110
Figure 7. Example of Sub-Factor Derivation, with Data of the Market Profile Subfactor

5.3 .................................................................................................................................... 111
Figure 8. The Expert Panel Survey Questions Based on the Propositions ..................... 115

INTRODUCTION

A range of organizations have addressed anti-counterfeit initiatives but there has
not been an overall academic review of the strategy with an integrated Counterfeit
Product Risk Model. This research focuses specifically on product counterfeiting and not
on currency, document forgery, piracy, or artwork forgery. A deeper understanding of
the strategic drivers behind counterfeiting will assist industry and agencies in making
more effective and efiicient decisions to deter the threat. These include the selection of
anti-counterfeit packaging components (e. g. holograms, RFID, color shifting inks, etc.)
based on an understanding of the cost and benefit relationships (Lukasik, 2008).

The product counterfeiting threat is increasing in risk and awareness (Fourtou,
2006; INTERPOL, 2007; Withington, 2005). Consumer and corporate awareness is
rising due to publicized events, such as the FDA Similac Infant Formula product recall,
the publicity from the book “Dangerous Doses,” the FDA Combating Counterfeit Drugs,
report the 18 million dose recall of the popular cholesterol drug Lipitor, the melamine
(counterfeit additive) in pet and human food, and the counterfeit branded toothpaste in
US dollar stores just to name a few (Eban, 2005; FDA, 2007b; Kontnik, 2004; Xin &
Stone, 2008).

Consumers have a growing awareness that product counterfeiting has moved
beyond luxury goods and that it occurs in developed countries such as the United States
and many others (Hopkins, Kontnik et al., 2003; Phillips, 2005). It has been emphasized
fi'equently that no product is too cheap to be counterfeited (FDA, 2006a; OECD, 2007b;

WHO, 2000).

The FDA Combating Counterfeit Drugs report specifically identifies the need for
a study in this area and for further development of a process to review and categorize the
risk. “FDA also plans to support the development of criteria that contribute to
counterfeiting risk. .. to assist stakeholders in focusing their use of anti-counterfeiting
technologies as effectively as possible” (FDA, 2004). In addition, FDA concerns
regarding the vulnerability of imported foods have brought a deeper focus to both the
counterfeit food threat and the broader economic food fraud concept.

Even with all of this risk awareness of the risk, there is little detailed public
information on counterfeiting specifics or risk models. There are many reasons for this,
including the fact that the information is fi'equently classified or confidential. Many other
risk research areas do not have a criminal element that warrants this level of sensitivity of
sharing information (e. g. packaging safety, food safety, financial markets, actuarial
tables, microbiological contamination, environmental exposure, equipment safety, etc.).

The overall goal of this research is to show that an anti-counterfeit strategy is
most efficiently and most completely reviewed from an academic perspective that is
based in the packaging discipline since most industry solutions are coordinated in the
packaging function by packaging managers. This goal is achieved through several key
objectives: review anti-counterfeit strategy, review risk analysis theory, review applicable
risk models, develop a Counterfeit Product Risk Model, review the findings with an

Expert Panel Survey, and test the model using an industry case study (Figure 1).

 

ummmm

Search Risk Concept

Research

     

Development

Risk Model
Research

Factor

Research

 

 

 

 

Case Study Survey

 

 

 

 

 

 

 

Figure 1. Development of the Risk Model Using a Multiple Component Research Approach
The objective of the research is to take a holistic, all-encompassing approach to
analyzing anti-counterfeit strategy, and specifically to develop, and theoretically support,

a Counterfeit Product Risk Model

1.0 LITERATURE REVIEW

This section will introduce the counterfeit threat, scale, and scope, including an
overview of regulations and countermeasures. This section intentionally presents a very
wide overview of the counterfeit threat and countermeasures, to present a broad view of
the topic, and to better assist in developing the prediction of a risk model. Particular
attention will be paid to the aspects of the counterfeit threat that apply to establishing the
root factors in the Counterfeit Product Risk Model: Counterfeit-History, Counterfeit-
Attractiveness, Counterfeit-Ability, Counterfeit-Hurdles, and Market Profile.

Prior to any strategic review of product counterfeiting, it is important to establish
some key concepts (N aim, 2005):

Illicit trade is driven by high profits, not low morals
Illicit trade is a political phenomenon - illicit traders cannot prosper
without help from governments or accomplices in key public offices.

0 Illicit trade is more about transactions than products - we are so

accustomed to parsing the illicit trades into separate product lines

o Illicit trade cannot exist without licit trade - all illicit businesses are

deeply intertwined with licit ones. Indeed, traffickers have strong

incentives to combine their illicit operations with legitimate business
ventures.

Illicit trade involves everyone - someone is buying...
Governments can’t do it alone

The US. Food and Drug Administration (FDA) and the World Health
Organization (WHO) both recognize counterfeit prescription drugs as a public health
threat, with the WHO going so far as to call it a global public health crisis (FDA, 2004;
WHO, 2007a). In a presentation at the World Health Organization’s IMPACT
(International Medical Products Anti-Counterfeiting Taskforce) Conference, the Council

of Europe Director General called counterfeiting a “silent pandemic” (Vladychenko,

2006). The US Federal Bureau of Investigation (FBI) has called counterfeiting “. . .the
crime of the 21St century” (Fourtou, 2006). Interpol has stated “Counterfeiting is
increasing relentlessly year on year” (W ithington, 2005), an “. . .the losses due to
transnational organized criminals to be “hundreds of billions of dollars annually”
(INTERPOL, 2007).

For brand owners versus the counterfeiters, “[often] the best they can hope for is
that the counterfeiters find them enough of a nuisance that they knock off another product

inst ” (Phillips, 2005).

1.1 Counterfeit Definition Overview

One challenge of reviewing the counterfeit risk and threat is a lack of agreement
on the most basic definitions. Different countries, organizations and industries apply a
range of definitions (Collins, 2004; Council of Europe, 2006; WHO, 2007b; World
Intellectual Property Organization, 2008). Reviewing the unique aspects of each of these
definitions is important in developing the Counterfeit Product Risk Model, to understand
how various factors influence the overall prediction of risk.

With such a wide range of definitions used for counterfeiting, it is worth
beginning with a review of the basic dictionary definitions of the key terms (Merriam-

Webster, 2004):

0 Counterfeit — “to copy, with the intent to deceive” and “made in imitation
of the genuine so as to deceive”
Diversion -- "a turning aside from a course, activity, or use
Simulation -- "1: the act or process of simulating 2: an object that is not
genuine 3: imitation by one system or process of the way in which
another system or process works" -- simulate ": to give or create the effect
or appearance of'

o Tamper -- "1 : to carry on underhand negotiations (as by bribery) <~with
a witness> 2 : to interfere so as to weaken or change for the worse <~ with
a document> 3 : to try foolish or dangerous experiments"

1.2 Counterfeit Definition Detail

Counterfeiting encompasses several key concepts. Many of the industries and
agencies involved with counterfeit products have published their own definitions. A
review of those definitions will be included in the following sub-sections.

The definition of the word counterfeit is evolving to cover both a macro, or
umbrella term for the entire concept, and a micro, or specific term for the copying of the
entire product and package, as is seen in the titles of several major reports by PIRA
International (Pira), the US Food and Drug Administration (FDA), and the Organization
for Economic Co-Operation and Development (OECD) (FDA, 2004; O'Kane &

Gopalkrishnan, 2005; OECD, 2007c).

1.2.1 Counterfeit

Counterfeiting is the copying of a product and/or package to deceive others into
believing that the product and/or package are/is genuine—also called a knock off or fake.
The FDA and WHO both includes deception about the source of a product in its
definition of counterfeiting (O'Kane & Gopalkrishnan, 2005; US Department of State,

2006; WHO, 2007a).

A more detailed review of the FDA and WHO definitions reveals important
details and similarities in their definitions.

The FDA definition of counterfeit drugs states:

US. law defines counterfeit drugs as those sold under a product name

without proper authorization. Counterfeiting can apply to both brand

name and generic products, where the identity of the source is deliberately

and fi'audulently mislabeled in a way that suggests that it is the authentic

approved product. Counterfeit products may include products without the

active ingredient, with an insufficient quantity of the active ingredient,

with the wrong active ingredient, or with fake packaging (FDA, 2007a).

The FDA clearly states “. . .sold. .. without proper authorization” in the definition
of counterfeiting, which would thus include unauthorized product diversion and parallel
trade as “counterfeit.”

The World Health Organization (WHO) definition of counterfeit medicines states:

A counterfeit medicine is one which is deliberately and fraudulently

mislabeled with respect to identity and/or source. Counterfeiting can apply

to both branded and generic products and counterfeit products may include

products with the correct ingredients or with the wrong ingredients,
without active ingredients, with insufficient active ingredients or with fake

packaging (WHO, 2007b).

The definitions used by the WHO and FDA are virtually identical, to the extent
that many of the same-and in some cases exact—words and phrases are used. A technical
difference is that the FDA uses “without proper authorization,” whereas the WHO
identifies a fiaudulent “source,” so WHO does not include diversion or parallel as
“counterfeit.”

Counterfeit products can be split into deceptive and non-deceptive product groups
(Green & Smith, 2002; Hopkins, Kontnik et al., 2003; Xuemei & Cleopatra, 2007).
Deceptive counterfeits are products that are placed into supply chains with the intent to

deceive the consumer into believing that the product is genuine in every way. Non-

deceptive counterfeits are products that do not try to deceive the consumer into believing
the products are genuine by their positioning in the market whether through the type of
retail outlet in which they are sold (flea market, etc.), the price (exponentially low), or the
quality (poor).

1.2.2 Diversion

Diversion is the distribution of a genuine product outside of its intended market; it
is also sometimes referred to as: parallel trading, gray market, smuggling, or product
arbitrage (Liang, 2006; O'Kane & Gopalkrishnan, 2005; Paradise, 1999; Phillips, 2005).
“Grey market, or parallel trade, may or may not be legal, depending on the country’s
intellectual property laws” (Bix, Clarke et al., 2007). To try to increase competition and
lower prices, countries such as Germany have explicit laws mandating importation of
drugs originally intended for consumption in other countries (Groupement International
de la Repartition Pharmaceutique, 2006).

Diversion and parallel trade have many enablers and another is transshipment
(also referred to as “trans-shipment” or “transhipment”), where the country or source of
origin can be disguised in the confusion of moving product through relatively (or
completely) unregulated and unmonitored facilities (European Commission, 2006;
O'Kane & Gopalkrishnan, 2005; OECD, 2007a; Union des Fabricants (on
INTERPOL.int), 2004; United Nations International Narcotics Control Board, 2006;
World Customs Organization, 2007c). In this case, a product fi'om a suspected
counterfeiting country may be sent to an intermediary country, repackaged and then
reshipped from that country to disguise the genuine origin (Bogdanich, 2007b; World

Customs Organization, 2005). This concept is also referred to as “neutralization,”

whereby a suspicious country of origin is hidden and replaced on the shipping paperwork
with a less suspect, or more neutral, country of origin (Bogdanich, 2007a; Rudolf &
Bernstein, 2004).

The gray market, which is comprised of authorized and secondary distributors
(who may be authorized for some products), has had a long history of adding competition
and efficiency to the pharmaceutical and healthcare supply chains, even though the
diverted or gray market products increase the opportunities to co-mingle fi'audulent and
legitimate products. (Council of Europe, 2006).

Product arbitrage—which essentially seeks to obtain lower prices for a market—is a
common, even respected, procurement best practice for many commodities such as oil
and cotton. Those engaging in the practice are sometimes called “Arbs” (deKieffer,

2006b)
1.2.3 Simulations

Look-a-Like products are also knows as mimics, knock-offs, or simulations, and
they attempt to confuse the consumer with similar brand names or packaging (FDA,
2003; Liang, 2006; Naim, 2005; O'Kane & G0palkrishnan, 2005). A Look-a-like product
does not duplicate the brand name. According to Webster’s Dictionary (yr), A
“knockoff” (noun) is an imitation of a product that does not attempt to duplicate the
brand name; to “knock ofi” (verb) a product is to make a copy[0] which may, or may not,
be an exact replica of the brand and design (Merriam-Webster, 2004). The design patent
law can be extremely complex to decipher, and the courts are fi'equently engaged to
define what is legal. In the packaged goods industry, design patents are more commonly

referred to as “trade dress” (Zaichkowsky, 2006). Trade dress is the “totality of

elements” which combines to create a type of trademark or identity (US Department of
State, 2006),,

1.2.4 Unauthorized Production

Unauthorized production is when a legitimate contract manufacturer produces a
product that was not authorized, and does so without the knowledge of the brand owner.
This could be due to over production in an authorized facility or contracted production
shifted to an unauthorized facility or supplier. This is also called an over-run, ghost shift,
fourth shift (manufacturing usually runs 3, eight-hour shifts), ghost plant, or ghost
supplier. (Booth, 1999; Council of Europe, 2006; Hopkins, Kontnik et al., 2003;
"Imitating property... ," 2003; Paradise, 1999; Parloff, 2006; Waite, 2004). The illegal
remanufacturing and sale of some wom-out parts can also be fall into this category
(Branch, 2008). Unauthorized overruns at an authorized facility is especially difficult to
prosecute, since all of the materials, design, and even the production equipment can be
“genuine.” The plaintiff must prove that this product violated the license (Booth, 1999;

Coalition Against Counterfeit and Piracy, 2006).
1.2.5 Tampering

“Product tampering can take one of three forms: grazing is the sampling of
products before purchase; pilfcring is theft during distribution or at a retail outlet;
malicious tampering is the intention to do damage or extort money” (O'Kane &
Gopalkrishnan, 2005). In terms of counterfeiting, tampering is considered small-scale; it
includes actions such as refilling used packages and returning to retailers for full credit,
or changing expired date codes (Collins, 2004; DiLonardo & Clarke, 1996; FDA, 2004;

Jotcham, 2005).

10

1.2.6 Theft

When considering counterfeit products, theft is an important concept to consider,
since the product was procured illegally and is often sold to unsuspecting consumers.
The product could have been mishandled or abused and thus become dangerous. Stolen
products can be co-mingled with genuine product, illegally refilled, shoplifted, pilfered,
hawked, boosted, and or taken through organized retail theft (ORT) (Chaudhry, 2006;
Grocery Manufacturer's Association, 2007; Kroft, 2004; National Retail Federation,
2006; Sodipo, 1997). This can include large scale theft, as in the case of cargo theft, or
the theft of entire tractor-trailers or seagoing shipping containers (Brekke, 2008). The
FDA would define this as a counterfeit product since the source would have been
(fisgunwd.

1.2.7 Piracy

Piracy is frequently mentioned in association with counterfeiting, but they are
distinctly different concepts (Balfour, 2005; European Commission, 2007 ; INTERPOL,
2007; Paradise, 1999; US Department of Commerce, 2007 ; US Department of State,
2006; World Trade Organization, 1994). Piracy refers to the “unauthorized reproduction
for commercial gain of literary, musical, artistic, and other copyright works, but may also
be used in some cases in the context of trademarked or patented works” (W allerstein,
Mogee et al., 1993). The EC 1383 guidance document clearly differentiates “pirated
goods” and counterfeits by defining piracy as violation of copyright (European Union,

2003).

11

1.3 Scale of the Threat

There are many reasons for the increase in scope and scale of the counterfeit
threat: globalization, the growth of brands, the growth of imaging and printing
technology, and organized crime shifting its focus and resources inthis direction
(Acheson, 2007a; Arrufiada & Vazquez, 2006; Closs & McGarrell, 2004; Hopkins,
Kontnik et al., 2003; Kelly, 2005; Naim, 2005; van Schendel & Abraham, 2005; Yar,
2005). The underlying driver of the counterfeit boom is the massive profit that can be
made by unscrupulous manufacturers, in addition to relatively low enforcement and
penalties for counterfeiting.

The Counterfeit Intelligence Bureau (CIB), of the International Chamber of
Commerce (ICC), estimates the worldwide trade in fakes to be 5-7% of the world’s trade
(Counterfeiting Intelligence Bureau, 1997, 2007). That estimate, which equates to
US$500-$600 billion annually, has been referenced by a wide range of other agencies and
organizations (European Commission, 2007; United States Department of Justice, 2006;
World Customs Organization, 2006). What is especially surprising is that only a fi'action
of total sales (estimated at 4-10% of all counterfeited goods), in an measurement are
believed to be in what would be considered “luxury goods” (Alcock, Chen et al., 2003;
Phillips, 2005). “In 2006, luxury goods accounted for only one percent of all the items
intercepted by the 27 Customs administrations of the European Union (EU)” (World
Customs Organization, 2007a).

The grth in counterfeiting is astounding, for example, the Center for Medicine
in the Public Interest (CMPI) has estimated that the worldwide counterfeit

pharmaceutical threat will grow from US$46 billion in 2005 to US$75 billion by 2010

12

(Schwarz & Wong, 2006). The counterfeit product market has an even faster growth
rate; it is now the size of the worldwide trade in illegal drugs (heroin, cocaine, marijuana,
etc.), which is estimated at ~US$500 billion (Budima, 2006; Union des Fabricants (on
INTERPOL.int), 2004; United Nations Development Programme, 1999). Criminal
organizations that expanded from illegal drugs to counterfeiting have, by definition,
essentially doubled their potential market.

Counterfeiting overlaps with other illicit trades, organized crime, and

terrorist networks. A natural connection, given the similarities in product

and format, is between the fake medicine and illegal drug trade, which
have come to follow many of the same routes (Naim, 2005).

The estimates for the impact of counterfeiting on the global marketplace are
elusive and the CIB states, “Given the clandestine nature of the business, such estimates
are no more than educated guesses” (Counterfeiting Intelligence Bureau, 1997). There
are a number of reasons for this: the clandestine nature of the counterfeiters, the reduction
in consumer detection of fakes due to increased quality, the low incentive for product
companies to divulge counterfeiting, and the fact that the core data collection comes only
from seized goods (Bialik, 2007a; Ivanova, 2007; World Customs Organization, 2007b).
Unfortunately, determining the level of counterfeiting is made more complex because the
dollar amounts are sometimes defined using the value of the genuine product (if sold at
genuine retail price) and sometimes by the value of the sale of fake products (sold as fake
product) (OECD, 2007c). PIRA’s Collins (2004) also states that: "Statistics in this area
are difficult to validate both because of the clandestine nature of the threats and some
brand owners' reluctance to talk, and the array of different authorities collecting statistics
for varying purposes." The estimates are given as a range, since the data is based on a

combination of estimating what percentage of product gets caught, and educated guesses

13

(heuristic perceptions) about what has passed through (Ivanova, 2007; Los Angeles
County Development Corporation, 2007). Regarding the customs seizure results, "We
wouldn't consider the seizure rate to be a random sample of the extent of counterfeit
goods coming in," says Loren Yagar, of the Government Accountability Office study (as
quoted in (Bialik, 2007b». To further compound the complexity of estimating a scale of
counterfeiting, a large amount of the work in the counterfeit area is either classified by
governments or considered confidential by companies.

The economic impact can also be measured in lost sales for companies and lost
tax revenue for governments (OECD, 2007b). The tax revenue loss can be in the billions,
even for US cities. Los Angeles estimates their annual losses to be $483 million (Los
Angeles County Development Corporation, 2007) and New York City estimates their
annual losses to be $1 billion (New York City Comptroller's Office, 2004).

Table 1 demonstrates how some counterfeit estimates are derived. For example,
the US Customs and Border Protection Bureau of the Department of Homeland Security
(US Orstoms) present actual data on the amount of product seized during a year — 2004
in this example. If US Customs stated that $138 million was seized, then to figure the
total estimated value of counterfeited goods, the analyst must estimate the percent of
product that was confiscated. This table then estimates the “implied value of counterfeit
goods” based on an “assumed interdiction rate (%).” So if $138 million represents 1% of
all counterfeit goods, then the “implied value of counterfeit goods is $13.8 billion, as the
table demonstrates (Los Angeles County Development Corporation, 2007). Of course,

the actual counterfeit value and the actual interdiction rate are unknown and unknowable.

14

Table 1. Estimated Value of Counterfeit Goods Based on Success Rate of Seizures

 

Estimated Value of Counterfeit Goods
Based on Success Rate of US. Customs and Border Protection Seizures, 2004
(Percentage Rate and Billions of Dollars)

 

US. Customs Assumed Interdiction Rate (%) 10.0% 5.0% 1.0% 0.2% 0.1%

 

 

Implied Value of Counterfeit Goods Entering the US. $1 .43 $2.83 $13.8B $69.18 51383

 

 

 

 

 

 

 

(Note: the report quotes $138. million in counterfeit goods seized by US Customs in 2004)

The US Customs and Border Protection Bureau reported that twenty-six million
commercial shipping containers were imported in 2008 (US Customs and Border Patrol
Bureau, 2008). The US. FDA estimates that they inspect less than 1% of all imported
food products, and that around 0.3% is actually lab tested (FDA, 2002; Paradise, 1999).
The Rotterdam shipping port, one of the busiest in the world, estimates they “inspect”
approximately 3% of the products that pass through their system (Phillips, 2005).
“Inspect” could mean anything from simply reviewing documentation to opening and
physical inspection of the shipping containers (European Union, 2003). The World
Customs Organization (W CO) has discussed the difficulty of determining fakes even
when the product is inspected (World Customs Organization, 2007a).

While quantitative determination of the counterfeit threat is impossible to
determine, the existing data points and anecdotal evidence emphasize that this is a serious
threat (Bialik, 2007a; United States Government Accountability Office, 2008a). It is
important to accept that the financial opportunity for the counterfeiters is astronomical,

which leads to great incentive and motivation to grow the scope and scale of the crime.

15

1.4 Scope of the Threat

The scope of the counterfeit product threat is widespread. Product counterfeiting
could be an economic threat, a public health threat, or both. It is important to note that
for the criminals doing the counterfeiting, the public health threat may be quite real but
occurs through negligence and not intent (PriceWaterhouseCooper, 2007; United States
General Accounting Office, 1995).

A general list of some counterfeited products is noted from the Global Business
Leaders Alliance Against Counterfeiting Global Business Leaders (GBLAAC) in Table 2

(Global Business Leaders Alliance Against Counterfeiting, 2003):

Table 2. List of Counterfeit Products from GBLAAC

 

 

 

 

 

 

 

 

 

 

 

 

Adhesive strips Agrochernicals Baby food

Batteries Canned meat and fish Cigarettes

Clothing and footwear Coffee Cosmetics and skin creams

Credit cards Diving masks Dry food gpods

Golf clubs Household appliances Household detergents and
cleaning products

Inkjet and toner cartridges Motor oils Parts for automobiles, trucks and
other machinery

Perfume Pharmaceuticals Photographic film

Razor Blades Rubbish bags Shampoo

Soft drinks Software, audio and video Sportswear

Super glue Tea Toothpaste and toothbrushes

Toys Wines and spirits Zippers

 

 

 

There are even reports of foodstuffs such as waffles, apples, candy and chewing
gum in Table 3 (European Commission, 2005). A WCO report lists some other

counterfeited products (World Customs Organization, 2007a).

Table 3. List of Counterfeit Products from WCO

 

 

 

 

 

 

 

Soup Mineral water

Breast implants Contact lenses

Pharmaceutical products for treating life- Pacemakers

threatening conditions such as breast cancer

and high blood pressure,

Weapons of war Automobile brake disc pads,
and even a WCO training course on CD

 

 

16

 

A random listing of the public health threats from the World Health Organization

(WHO) shows the following (WHO, 2007a):

e 89 children died in Haiti in 1995 and 30 infants died in India in 1998
due to the consumption of paracetamol cough syrup prepared with
diethylene glycol (a toxic chemical used in antifreeze).

o In 2001, in South-East Asia, a Wellcome Trust study revealed that
38% of 104 anti-malarial drugs on sale in pharmacies did not contain
any active ingredients.

0 In Cambodia, in 1999, at least 30 people died after taking counterfeit
anti-malarials prepared with sulphadoxine—pyrimethamine (an older,
less effective anti-malarial) which were sold as artesunate.

All countries, even the biggest source economies, are adversely impacted by
product counterfeiting. It is estimated that 250,000 Chinese die each year from
counterfeit or “substandar ” medicines though this does not account for all counterfeit
products, so the actual numbers may be even more staggering (Morris & Stevens, 2006;

Schwarz & Wong, 2006).

1.5 Counterfeit Drugs

Counterfeit drugs will be covered in more detail than other product segments
because there are more regulations and public reports available on this industry segment.
The terms pharmaceutical, prescription drug, drug, and medicines are often used
interchangeably in the literature, and thus, in this section (Council of Europe, 2006).
Healthcare products (such as bandages and surgical gloves) and medical devices (such as

pacemakers and intravenous bags) are usually clearly separated from pharmaceuticals.

1.5.1 The Counterfeit Drug Scale

As noted in a New England Journal of Medicine article, the FDA Director of the
Counterfeit Drug Task Force estimated that “less than 1%” of US. prescription drugs

were counterfeit (Rudolf & Bernstein, 2004), but with $216 billion spent on 3.4 billion

17

prescriptions, this would still equate to $2.6 billion worth of counterfeits and 34 million
counterfeit prescriptions. The FDA estimates do not include legitimate or counterfeit
prescription drugs that are sold on the “street” by a traditional “drug dealer.”

The CMPI has estimated that the worldwide counterfeit pharmaceutical threat will
grow fi'om $46 billion in 2005 to $75 billion by 2010 (Schwarz & Wong, 2006). This
estimate is fi'equently referenced by many organizations such as the EU, WHO, FBI, and
FDA (Collins, 2004).

PIRA has developed estimates that the market share of counterfeits are (Collins,
2004):

5% in North America,
5.8% in Europe,

10.5% for the rest of the world, and
an overall global rate of 6%

A World Health Organization survey of counterfeit prescription drugs estimates
that 1-10% are counterfeits in developing countries, that 40% of counterfeit incidents are
in developed countries; that developed countries with effective regulatory systems had
fewer incidents; and that over 50% of all internet purchases were counterfeit (WHO,
2007a).

The FDA Combating Counterfeit Drugs report is a summary of industry
perspectives and surveys, and it is not meant as guidance nor is it an official position
document. The FDA report does list attributes that can contribute to products likely to be
counterfeited (FDA, 2004). The factors listed here, and the additional detail provided
later, are a mixture of probability and severity factors that make up the overall risk

(which is a combination of risk and severity) (FDA, 2004).

18

Drugs history of counterfeiting;

Drugs price;

Drugs volume;

Drugs dosage form;

Drugs clinical uses; and

Whether similar products had a history of being counterfeited.

As was stated earlier, the discussion and data relating to counterfeiting is often

confusing since different organizations or agencies use different base definitions.

1.5.2 The Counterfeit Drug Scope

The World Health Organization (WHO) investigated seized counterfeit
prescription drugs and published their insights. They focused their analysis on the
physical product (WHO, 2007b).

“In terms of types of counterfeits and their magnitude, the products reported can be
grouped into six categories:

Products without active ingredients, 32.1%;

Products with incorrect quantities of active ingredients, 20.2%;

Products with wrong ingredients, 21.4%;

Products with correct quantities of active ingredients but with fake packaging,
15.6%;

Copies of an original product, 1%; and

0 Products with high levels of impurities and contaminants, 8.5%.

In terms of a public health threat, several authors corroborate the results of the
WHO survey which revealed that most counterfeiting does not pose an active threat, such
as anthrax or poison or a bomb (Council of Europe, 2006; OECD, 2007d; WHO, 2007b).
The WHO survey reported that 91.5% of counterfeit drugs were a placebo at worst
though a placebo cholesterol control medicine, vaccine, or birth control pill can create a
public health dilemma or threat (WHO, 2007b). Of the 8.5% of products that contained
some threat, many included contaminants derived from poor manufacturing practices,

such as bacteria. With the Heparin counterfeiting in 2007, the public health threat was

19

from an allergic reaction to a contaminant that cost one-twentieth of the price of the
genuine ingredient (Bogdanich, 2008). The COE report identified these as either active
or passive threats (Council of Europe, 2006).

Another threat associated with counterfeit prescription drugs mentioned by the
World Health Organization was in the undermining of consumer confidence in a novel
treatment or in the drug itself (“pharmacovigilance”) (WHO, 2006) such as happened
with anti-malarial drugs, which have a recorded 10-70% counterfeit rate (WHO, 2007a).
The WHO noted a 1995 incident in Niger, in which 50,000 people were inoculated with a
counterfeit vaccine that was donated by another country. The counterfeit vaccines
resulted in 2,500 deaths (WHO, 2007a). This is a specific example where patients lost
confidence in the healthcare system and the very drugs that could be helping them.

In other instances, the consumption of wrong or ineffective drugs could contribute
to antibiotic resistance of bacterial pathogens, leading to diseases that are harder to treat
(Liang, 2006).

These are a few examples of counterfeiting from the FDA report on counterfeit
prescription drugs (FDA, 2003):

o Procrit, a drug used by cancer and AIDS patients, was recently

counterfeited and the drug was replaced with non-sterile tap water,
which could have caused a severe infection of the bloodstream.

o In another recent counterfeiting incident, white tablets with “aspirin”

imprinted on them replaced Zyprexa, a drug used for schizophrenia
and acute bipolar disorder.

- Counterfeiters also have been known to switch a lower-strength drug
for [a] higher strength drug. As a result, patients receive lower than
expected doses of [the] drug, leading to ineffective treatment and
therapeutic failure.

20

1.6 Countermeasure Action and Deterrence

Agencies and organizations, such as the FDA and the WHO, have begun
programs to review and deter drug counterfeiting. These include focused programs such
as the FDA Counterfeit Drug Task Force. Also, the WHO has implemented the

International Medical Products Anti-Counterfeiting Task force (IMPACT).

1.6.1 Countermeasure Regulation and Enforcement

Despite the often low level of enforcement or prosecution, it is important to
emphasize the regulation and enforcement of intellectual property rights (United States
Government Accountability Office, 2008a). Without regulation (deterrent laws),

regardless of the actual level of enforcement, counterfeiting would not be illegal.

1.6.2 Intellectual Property Rights

Counterfeiting is illegal in most countries since the practice violates intellectual
property right laws of trademark, or design infringement, or both. Intellectual property
rights (IPR) are protected under the “Agreement on Trade Related Aspects of Intellectual
Property Rights, Including Trade in Counterfeit Goods,” also known as “TRIPS” (World
Trade Organization, 1994).

In the United States, the United States Code (U SC) is “the law” and the Code of
Federal Regulations (CFR) is “how to implement the law,” including enforcement,
reporting, and punishment (United States Patent and Trademark Office, 2008).
Prosecution against infiingers can be initiated by the government or by citizens, and the
action can either be under criminal law (offense against the state) or civil law (offense

between two citizens) (US Department of State, 2006). The “Enforcement of Intellectual

21

Property Rights Act of 2008” (PRO-1P Act) allows the US Department of Justice to bring
civil cases against infiingers, which is significant since the burden of proof is lower in
civil cases than in criminal cases (Leahy, 2008).

There are several milestones in the evolution of counterfeiting laws (Marvel,
2008):

o Trademark Counterfeiting Act of 1984
o Anticounterfeiting Protection Act of 1996
0 Stop Counterfeiting in Manufactured Goods Act of 2006

The US Department of Commerce has an interdepartmental initiative on
intellectual property rights at StopFakes. gov, which identifies the most significant
intellectual property rights’ provisions (US Department of Commerce, 2007).

The counterfeit trademark crime is set out at 18 U.S.C. 2320 ;
Criminal infiingement of copyrighted works is set out at 17 U.S.C.
506(a) and 18 U.S.C. 2319;

The counterfeit labeling provision is set out at 18 U.S.C. 2318;
Theft of trade secrets [is] prohibited by 18 U.S.C. 1831 and 1832

In addition, the Digital Millennium Copyright Act (DCMA), at 17 USC 1201-
1204, is applied for copyright of recordings or software (US Department of State, 2006).

There are some civil (rather than criminal) laws, such as 15 USC 1117 and 15
USC 1536, which focus on specific remedies for the plaintiffs and define the authority of
the US Customs (Marvel, 2008).

In general, the laws continue to be adapted to be more efficient in closing gaps to
more appropriately address the overall counterfeit threat. For example, 21 USC 331 (i2-
i3) makes it illegal to possess the tools and mechanisms to make the counterfeit drugs.
That being said, 18 USC 2320 is used more often since it “has a bigger stick” of

increased penalties (Marvel, 2008).

22

The Intellectual Property Rights (IPR) laws cover trademark, patent, copyright,
and trade secrets. The basic laws that govern the counterfeiting are for trademarks and
patents, and are usually present in the same violation. Specifically a trademark covers a
word/design/symbol (a brand logo), whereas a patent covers a utility/design/plant (a
design of a machine). Most countries have similar laws, and there are many international
treaties on the subject.

The Anti-Counterfeiting Consumer Protection Act has applied the RICO Act (The
Racketeer Influenced and Corrupt Organizations Act) to counterfeit crimes (18 USC Sec.
1963, 2006). The RICO act was successful in the 1980’s when it was applied to illegal
narcotics trafficking (e. g. cocaine) in the US. The Act’s strongest feature is the seizure of
“real property, including things growing on, affixed to, and found in land; and tangible
and intangible personal property, including rights, privileges, interests, claims and
securities” (18 USC Sec. 1963, 2006).

A key to the RICO Act in the US is that, when applied, it has very sharp teeth.

If you catch them and don’t prosecute, it breaks down. If you prosecute

but don’t sentence, it breaks down. If you sentence but don’t take their

money, you’ll have a problem. So what is very important is that you need

to engage the full spectrum of criminal justice (Phillips, 2005).

Nevertheless, actual prosecution of formal counterfeiting activities is difficult
since two “states of mind” need to be established: intentional trafficking in counterfeits
and knowing use of a counterfeited mark (Marvel, 2008). Regarding the role of diversion
in counterfeiting, the international intellectual property rights law firm, “Torys,” states
that every counterfeiter they brought to court claimed to be a secondary or gray marketer

selling genuine products outside of authorized channels, and not a counterfeiter

(Coalition Against Counterfeit and Piracy, 2006). If the plaintiff can convince the court

23

that the intent was diversion, this eliminates the application of the counterfeit laws
(Marvel, 2008).
There is a continued agency and industry effort to coordinate actions to

effectively deter counterfeiters (Lukasik, 2008).

1.6.2.1 Civil Liability for the Brand Owner

Foreseeable hazard, also known as willful negligence or willfirl blindness, is
being applied to counterfeit products (Coalition Against Counterfeit and Piracy, 2006;
Greenberg, 2003; Supreme Court of the State of Florida, 2003; University of Florida,
2004; Yankus, 2006). “Willful blindness” only applies when it can be proven that a
person “must suspect wrongdoing and deliberately fail to investigate” (Marvel, 2008). A
new type of “foreseeable hazard” for companies is included in the reporting requirements
of the Sarbanes-Oxley Act of 2002, which require the reporting of business risks, which
can extend to counterfeiting and diversion (Coalition Against Counterfeit and Piracy,
2006; Millar & Yeager, 2007; O'Kane & Gopalkrishnan, 2005).

1.6.2.2 Control of the Distribution of Products and Brands

Another legal aspect that applies to counterfeiting is the specific definition of
when, and to what extent, ownership and control of a product passes fiom the seller to the
purchaser. This is important to this research in terms of the risks inherent in diverted
product. The control of a diverted product is based on a country’s position on the
Exhaustion Law—also called the First Sale Doctrine covered under law (18 USC 109) (US
Department of State, 2006)—since the intellectual property rights owner exhausts the
control of distribution of a product upon the first sale of the product (Y eung & Mok,

2006). There are two perspectives on this law: after the brand has been put into trade

24

with the consent of the trademark owner, either they have control of the distribution of
their brand or they do not, they exhaust their control once the product is sold (Prahl,
2007)

To supplement the First Sale Doctrine, the United States uses a unique
interpretation of the Tariff Act (19 USC 1526) to block the import of a trademarked
product without the trademark owner’s approval (Tritter, 1989). This application helps
narrow a supply chain gap, and provides insight into the value of other counterfeit
countermeasures.

Many other global threats are being addressed with processes or systems that will
also assist in fighting counterfeiting, such as actions that increase the traceability of
international shipping containers, electronic pedigrees (history of ownership) for
prescription drugs, and the traceability regulations and processes for food.

1.6.2.3 Pedigree

The FDA has been pursuing a drug pedigree to authenticate products but a lack of
pedigree does not automatically indicate that a product is counterfeit (FDA, 2006b). “A
drug pedigree is a statement of origin that identifies each prior sale, purchase, or trade of
a drug, including the dates of those transactions and the names and addresses of all
parties to them” (FDA, 2006b). The drug pedigree regulations were first introduced in
1999, but have faced continual delays or injunctions as recently as 2008.

In the US, medical devices are implementing a method called “Unique Device

Identifiers” (CDRH/FDA, 2009).

25

1.6.3 Packaging Countermeasures

The Intellectual Property Rights holders (e. g. brand owners) face a tremendous
financial liability if their products are counterfeited, in loss of sales, cost of recalls, loss
of brand equity, and civil litigation. Brand owners have been successfully sued for not
protecting the marketplace from counterfeit products, even though they did not
manufacture the counterfeit product themselves (Coalition Against Counterfeit and
Piracy, 2006; Greenberg, 2003).

Anti-counterfeit package technologies take several forms and the technologies fall
into three general groups: packaging components, authentication, and traceability (FDA,
2004). The FDA Combating Counterfeit Drugs report categorizes packaging components
as authentication (the ability to prove genuine) technologies (FDA, 2003).) One product
or feature could fall into several categories. The packaging technologies can also be
deterrents by implementing “counterfeit-evident” and “counterfeit-resistant” actions
(Andel, 2006). By the same token, one anti-counterfeit feature could fall into several
categories based on its role in deterrence.

Of the anti-counterfeit packaging component technologies, there are three main
concepts: overt, covert, and forensic (which is also sometimes called reserved) (Collins,
2004; O'Kane & Gopalkrishnan, 2005). The FDA definitions are provided below (FDA,
2003):

o Overt technologies are protective measures that are easily visible to the

eye, such as holograms, color shifting inks, and some watermarks.

0 Covert technologies are protective measures that are not visible to the

eye and frequently require special equipment for visualization (and

authentication). These include some watermarks, certain inks and dyes
that fluoresce or absorb ultraviolet light, and invisible bar codes.

26

o Forensic technologies are protective measures that require
sophisticated analytical equipment, usually found in a forensic
chemistry lab, in order to be identified. These include chemical
markers, taggants, and other unique chemical properties of a
substance.

Overt technologies could be as simple as a complex package shape, a unique
design feature, or a unique component such as a child-resistant or tamper-evident feature.

Traceability, also referred to as “track and trace,” technologies include radio-
fi'equency identification (RFID) and barcodes, and are implemented to identify and
follow a product through the supply chain (Bix, Clarke et al., 2007; Dietrich, Puskar et
al., 2006; FDA, 2003; Greenberg, 2003).

Authentication includes features, processes and procedures that serve two
purposes: they make it harder to counterfeit a product and they provide a way to
determine if a product is authentic (FDA, 2004). Examples of authentication
technologies include: “color shifting inks, holograms, fingerprints, taggants, or chemical
markers embedded in a drug or its label” (FDA, 2004). Other features could include
RFID tags, 2-D bar codes, electronic pedigree, and even just more closely monitoring the
supply chain.

A discussion of deterrent actions is not complete without considering the response
by the counterfeiters (Lukasik, 2008). “And counterfeiters are skilled at duplicating
holograms, ‘smart’ chips and other security devices intended to distinguish fakes from
the genuine article” (Balfour, 2005). Global marketing directors have stated that their
sophisticated, multi-year researched technologies have been knocked off very quickly
(Balfour, 2005; Collins, 2004). Even RFID is mentioned as “promising but not risk

proof” by both the Center for Medicine in the Public Interest and the FDA (FDA, 2003;

Schwarz & Wong, 2006).

27

A layered approach, combining several varied components and systems, is usually

implemented for the greatest impact.

1.7 Summary

Product Counterfeiting is clearly not a victimless crime. Counterfeit products are
a clearly recognized and significant economic and public health threat. The counterfeit
product threat is growing in scope and scale. An interdisciplinary, multi-layered
approach is necessary to curb the growth in counterfeits and the risk to the public.

The lack of specific public research and refereed publications underscores the
value of initiating counterfeit-specific academic research. The quantitative data on the
overall threat is very sparse and has been described as unknowable. Counterfeit product
estimates are a combination of actual seizures and educated guesses.

Strategic research and advanced models are called for but are not yet published, at
least not in the non-classified and non-confidential, open source literature. It is evident

that a strategic approach is necessary to effectively and efficiently deter counterfeiting.

28

2.0 RESEARCH ANALYSIS

This chapter analyzes three main areas that apply to developing the Counterfeit
Product Risk Model: (1) “Risk Factor Selection Analysisml” Research which explores and
develops detailed attributes of a counterfeit threat, (2) “Risk Analysis” Research which
explores the aspects and methods of risk analysis that apply to the model, and (3) “Risk
Model” Research which explores a wide-review of risk models. This chapter takes a
purposeful research based analysis the basics of counterfeiting, risk analysis, and risk
models to specifically address what should, and what should not, apply to the Counterfeit
Product Risk Model.

This section expands on the literature analysis to explore and develop key, root
factors that address the counterfeit threat and can be used in the development of the
Counterfeit Product Risk Model. Agency reports—specifically the FDA Combating
Counterfeit Drugs report—stress the need to have a counterfeit risk model that would
help brand owners review aspects of the counterfeit risk and prioritize the highest risk

products for deterrent actions (FDA, 2004).

29

 

Five factors were selected for the Counterfeit Product Risk Model, based on
literature and report reviews, and peer consultation.
0 Counterfeit-History: evidence of past counterfeiting
o Counterfeit-Ability: ease of mimicking genuine product
0 Counterfeit-Attractiveness: scale of fi'aud in relation to costs
0 Counterfeit-Hurdles: deterrent activities or systems
0 Market Profile: aspects of the consumer, the sales channel, the manufacturing

system, etc.

2.1 Risk Factor Selection Analysis

Sodipo (1997) presented one of the most complete and detailed discussions of the
attributes impacting counterfeiting, based on research related to the TRIPS (Agreement
on Trade Related Aspects of Intellectual Property Rights) agreement. Sodipo’s Key

counterfeit attributes were:

Profits

Cheap to copy

Easy to copy

Unsatisfied market demands
Difficulties of detection and proof
Non-deterrent laws and lacunae in laws
Poor government policies

Location of countries [production])

From a counterfeit product supply-side perspective, there are an additional set of
more detailed sub-factors to be considered from the driving factors of counterfeit supply

The Organization for Economic Co-Operation and Development (OECD) listed,
from a supply-side perspective a set of detailed factors that should be considered as

potential counterfeit drivers (OECD, 2007a):

30

0 Market Characteristics

High unit profitability
Large potential market size
Genuine brand power

0 Product, Distribution and Technology

Moderate need for investments
Moderate technology requirements
Unproblernatic distribution and sales
High ability to conceal operations
Easy to deceive consumers

0 Institutional Characteristics

Low risk of discovery

Legal and regulatory fiamework
Weak enforcement
Non-deterrent penalties

Also, the OECD listed, from a demand-side perspective a set of detailed factors

that should be considered as potential counterfeit drivers (OECD, 2007a):

0 Product Characteristics

Low prices
Acceptable perceived quality
Ability to conceal [fake] status

0 Consumer Characteristics

No health concerns
No safety concerns
Personal budget constraint
Low regard for IPR

0 Institutional Characteristics

2.1.1

Low risk of discovery and prosecution
Weak or no penalties

Availability and ease of acquisition
Socio-economic factors

Counterfeit-History

It has been emphasized repeatedly in the literature that the strongest indicator of

the risk of product counterfeiting was past evidence of product counterfeiting (FDA,

2004; Hopkins, Kontnik et al., 2003; O'Kane & Gopalkrishnan, 2005; OECD, 2007b;

Phillips, 2005). For example, the OECD’s Economic hnpact of Counterfeiting and

31

Piracy report specifically identifies the first of three steps in evaluating counterfeit
magnitude as “identification of those goods that have been detected in international trade
as counterfeit or pirated” (OECD, 2007c).

That being said, the opposite is not true—a lack of evidence of past counterfeiting
is not indicative of a low risk of future counterfeiting. The attacks can be random, a very
well protected product may get knocked off and a completely unprotected product may
never get knocked off (deKieffer, 2006a; FDA, 2003). This factor is important [0]since
the past counterfeiting or diversion is indicative of the potential for more counterfeiting

and diversion.

2.1.2 Counterfeit-Ability

“Counterfeit-Ability” is a measure of the challenge of duping consumers (or
retailers) into thinking they are buying the genuine product (deceptive counterfeits), or to
just make the counterfeit product sale (non-deceptive counterfeits). This factor includes
consideration of the product itself, the packaging, and even the authentication systems,
such as warranty registration (Collins, 2004; Global Business Leaders Alliance Against
Counterfeiting, 2003; Kontnik, 2004; OECD, 2007b; Sodipo, 1997; World Customs
Organization, 2007a). This factor [0]is important since the ability to actually sell a

counterfeit product is a major driver for counterfeiters.

2.1.3 Counterfeit-Attractiveness

Counterfeit-Attractiveness” is a measure of the ability to profit, taking into
consideration the costs of production and of getting products to, and into, a market

(Global Business Leaders Alliance Against Counterfeiting, 2003; Kontnik, 2004; OECD,

32

'2‘
in.
l

IBC \

I“?

2007b; Sodipo, 1997; World Customs Organization, 2007 a). A very large market size
may not always be a positive, if it contains very educated consumers (who would
recognize minor variations between the genuine and fake product) or if there are very
stringent deterrents in place (Lukasik, 2008). A very small market size may not always
be a negative, if the set-up and execution costs and hurdles are low. This factor [0]is

important since the ability to make a financial profit is a major driver for counterfeiters.

2.1.4 Counterfeit-Hurdles

“Counterfeit-Hurdles” is a measure that includes deterrent activities—whether
designed or inherent in a system—that create challenges to any of the other counterfeit
factors (Collins, 2004; Global Business Leaders Alliance Against Counterfeiting, 2003;
Kontnik, 2004; OECD, 2007b; Sodipo, 1997; World Customs Organization, 2007a). The
easiest hurdles for a brand owner to implement are packaging components, since they
only take one management decision, versus integrated supply chain or enforcement
solutions (McNeely, 2007, 2008). A hurdle could be a routine lab inspection of incoming
raw materials, which would increase the chance of a counterfeit additive being rooted
out. A characteristic of a supply chain that could reduce the impact of counterfeiting is
establishing exclusive regional licensing agreements, where the local company pursues
local enforcement of the brand (Kennedy, 2005; Kontnik, 2006; Studler, 2008). This
factor [0]is important since fewer barriers or hurdles equate to a higher Counterfeit-
Ability or Counterfeit-Attractiveness, but Counterfeit-Hurdles are unique and should be
considered in their own factor.

For the Counterfeit Product Risk Model, which is discussed in detail in Chapter 3,

if a product has many or complex hurdles, then the risk ranking would be “Low” for the

33

contribution of this factor to the overall risk of product counterfeiting. This process was

utilized for continuity in the model (low always being good).

2.1.5 Market Profile

“Market Profile” is a measure of the infinite aspects of product development,
production, supply chain specifics, retailing, consumer demographics, and geographic
marketplace dynamics that contribute to the Counterfeit Product Risk Model. This
f[0]actor is important since it considers characteristics that impact the counterfeit risk that

are not included in the other[0] factors.

2.1.6 Counterfeit Risk Sub-Factors

The literature and reports emphasized the value of breaking overall risk into
components, such as probability and severity. The components of probability and
severity have been broken down into qualitative characteristics such as high, medium,
and low (Todd, 2008). Further, it was recommended that the overall components, such as
probability, be broken down into unique factors and sub-factors.

In many cases, key attributes play a much more significant role than other
attributes, and the specific model sub-factors should be chosen by a specific company’s
expert panel (Aven, 2003; Cruz, 2002; Todd, 2008). The literature and reports that were
reviewed discussed some of the market factors in detail, for example, in the OECD report
that covered specific industries, a common set of sub-factors was not found that would
cover all products (OECD, 2007c). The identification and application of sub-factors that
are unique to each product or company are explored more in the Model Development

Section.

34

2.1.7 Risk Factor Selection Analysis Summary

In this Risk Factor Analysis Section, five factors that comprise the counterfeit
risks: Counterfeit-History, Counterfeit-Attractiveness, Counterfeit-Ability, Counterfeit-
Hurdles, and Market Profile were selected based on the information covered. The

research supported using, but not explicitly universally defining, sub-factors.

2.2 Risk Concept Research

Since product counterfeiting is a recently emerging and rapidly evolving risk, this
section will focus mainly on the risk assessment aspect of the overall risk analysis
concept. It is important to research the risk concepts to gain a perspective on the
components that apply to the Counterfeit Product Risk Model, and to firrther theoretically
justify what is, and what is not, proposed. This section reviews risk analysis and its three
core components of risk assessment, risk management, and risk communication. This
section will include a discussion of why classical statistical methods in general, cannot be
applied to small and incomplete data sets likely to be available to be used in the
Counterfeit Product Risk Model. Dr. Buchanan, a CFSAN/F DA risk assessment expert
offers a very useful overall perspective:

A good risk assessment should be sufficient to answer the question that is

being posed, but not so complex that it substantially delays or confuses the

need by risk managers, the users of the risk assessment, to reach a decision
(Buchanan, 2004).

Risk analysis, and particularly the subroutine of risk assessment and its antecedent

hazard identification, is a key aspect in developing an anti-counterfeit strategy.

35

.9

CO

in

Hz»

[5'
f I I

2.2.1 Overview of Risk Analysis

Risk is the exposure to a chance of loss (Claycamp, 2003; Kaplan, 1997). Risk
analysis is not a number or value in and of itself; it is a combination of three basic
concerns: what could happen, how likely it is to happen, and what the consequences are if
it does happen (CF SAN/FDA, 2002, 2003; Kaplan, 1997).

The goal of risk analysis, as with corporate quality control programs, is to reduce
errors, incidents, or risks as low as possible, but the total elimination of risk to a
probability of zero is usually not practical, or even desirable considering that there is a
point of diminishing return on the incremental risk mitigation investment (Banks, 2005;
Busta & Chaisson, 1997; Nakayachi, 1998; United States Government Accountability
Office, 2008b).

At the 2008 International Association for Food Protection (IAFP) Annual
Conference, Dr. Buchnan stated, “The concept of zero tolerance expresses an absolute
but has no meaning scientifically.” (Buchanan, 2008).

The goal of a process or system might be “zero defects” but the probability of a
defect would not be zero. With food products, even with the worst of substances, there is
some low level of tolerance before it is considered a public health threat (“threshold of
regulation” or TOR) (CFSAN/F DA, 2005). That being said, the presence of a
contaminant, even within acceptable tolerances, may lead a company to action such as a
recall. The process does try to reduce the magnitude of each error, to the point that the
errors are not big enough to result in a “defect.” There will still be variations, but they

are reduced to a point where they are not critical (Lipscomb & Lewis, 2004).

36

 

Risk analysis consists of risk assessment (and the antecedent of hazard

identification), risk management, and risk communication (see Figure 2).

Risk Assessment Risk
. Including Hmrd Management
Identification

       
 

Risk
Analysis

Risk
Communication

   

Figure 2. The Risk Analysis Concept and its Components

The FDA/CFSAN defines risk analysis this way: “Risk analysis is a tool to
enhance the scientific basis of regulatory decisions. It includes risk assessment, risk
management and risk communication activities” (CFSAN/F DA, 2002). Risk Assessment
is broken down to hazard identification, hazard characterization, exposure assessment,
and risk characterization. These are defined by FDA and USDA as:

0 Hazard Identification: “First, risk assessors and risk managers must
clarify the public health hazard that is the subject of the assessment
and any possible policy options that are under consideration”
(FSIS/USDA, 2003b).

0 Hazard characterization: “Next, the risk assessors must evaluate the
adverse health effects caused by the public health hazar ”
(FSIS/USDA, 2003b).

0 lllness/ death
0 Economic
0 Social/ political

37

0 Exposure Assessment: “Then, an exposure assessment must be
conducted to estimate the likelihood that the hazard will be present in
food, and if present, at what level” (F SIS/U SDA, 2003b).

0 Risk characterization: “Next, a dose-response model is constructed
to figure out at what dose or concentration that hazard will cause
illness or death” (FSIS/USDA, 2003b).

The FDA/CF SAN describes risk management and risk communication this way:
0 Risk Management: “The risk management phase involves using all of
the information gathered during the assessment to evaluate policy
options” (CFSAN/F DA, 2003).
0 Risk Communication: “Risk communication not only refers to
communicating the results of the risk analysis to the general public,
but also to the ongoing communication among risk assessors,

managers, scientists, regulators, and various stakeholders during the
entire process” (CFSAN/F DA, 2003).

It is important to clarify that throughout the literature, the terms hazard
identification and risk identification are often used for the same step in the process. This
step is sometimes considered as a separate precursor to risk assessment, leading to four
main steps to risk analysis, instead of three (Cervone, 2006; F SIS/U SDA, 2003b).

There is a great deal of research and literature in the area of risk assessment, and.
although there are methods of reviewing risk that are very quantitative and rigorous, what
is important to the Counterfeit Product Risk Model is that the process of developing
models or tools is recognized as a very fluid and creative art (Broder, 2000; Etherton,

2007; Jablonowski, 1996; Kaplan, 1997; Langford, 2002; Taylor, 2006).

2.2.2 Risk Assessment and Hazard Identification

A hazard is “Any real or potential condition that can cause injury, illness, or death
to personnel; damage to or loss of a system, equipment or property; or damage to the
environment” (Mil-Std-882D, 2000). Processes such as HACCP are excellent at

managing risks, once the risk has been identified. The very first and most basic step is

38

figuring which risks or hazards should be included in a full risk analysis (F SIS/USDA,
2003b). Deciding which hazards to actually consider is based on many factors, such as
familiarity and an assumption of the ability to actually mitigate the risk (Marvin, Kleter et
al., 2009). Research shows that in the period leading up to the financial market meltdown
of 2008, financial risk managers foresaw the dangers of the sub-prime market, but the
“meltdown” was considered very improbable and assumed to be a risk that should be
accepted since it seemed too big for any company to mitigate (Gerardi, Lehnert et al.,
2008) . Some companies, such as Goldman-Saks, did take action and reduced their
exposure to the sub-prime market even though it reduced their short-term financial
returns during a market boom (Kiviat, 2008). A danger is that the counterfeit risk hazard
may not be completely factored into the risk assessment since it could be considered
unknown or unmanageable, or a very remote possibility.

A detailed discussion on risk assessment is available from the United States
Office of Management and Budget (OMB) (Office of Management and Budget, 2006).

Types of Risk Assessments - Risk assessment is a broad term that

encompasses a variety of analytic techniques that are used in different

situations, depending upon the nature of the hazard, the available data, and

needs of decision makers. The different techniques were developed by

specialists from many disciplines, including toxicology, epidemiology,

medicine, chemistry, biology, engineering, physics, statistics, management

science, economics and the social sciences. Most risk assessments are
performed by teams of specialists representing multiple disciplines.

2.2.3 Probabilistic Risk Analysis (PRA)

Some risk, probability, mathematical, and quality control theories intertwine. A
review of “probabilistic risk assessment” highlights ways in which these areas are, and
are not, similar. Probabilistic risk analysis/assessment (PRA) is a highly regarded

method or tool that is part of the general risk analysis concept. While this provides

39

general insights, strict adherence to PRA does not apply directly to the Counterfeit
Product Risk Model.

Probabilistic risk assessment (PRA) is a method of quantifying risk analysis. The
general steps include: system identification, event identification, frequency estimation,
consequence analysis, risk estimation (frequency and consequence estimation), and risk
assessment (Zambrano, Sublette et al., 2007).

As stated in a NASA Newsletter, PRA is quantitative and by definition focuses on
“every life-cycle aspect of a complex engineered technological entity” (Starnatelatos,
2000). In addition, the US Office of Management and Budget’s “Proposed Risk
Assessment Bulletin” (2006) emphasizes using PRA for physical systems, such as “dams,
nuclear and chemical plants, liquefied natural gas terminals, space shuttles and other
physical systems” (Office of Management and Budget, 2006).

In the article “Using Probabilistic Risk Analysis to hnprove Risk Management,"
which appeared in the peer-review journal Risk Analysis, the benefits and limitations of
PRA were discussed as being “imperfect but useful” (J ablonowski, 1996). The article
emphasizes that PRA, and all risk analysis, is based on imperfect knowledge, so the
uncertainty must be commrmicated. In more uncertain events, conveying the
probabilities with words rather than numbers is a way to express greater uncertainty.

For the Counterfeit Product Risk Model, PRA has limitations since there is little
actual data on the frequency or severity of the loss. A formal PRA focuses on assigning
probabilities to a loss and bases the full analysis on historical data.

A formal PRA is also focused on discemable, discrete events rather than system-

wide, emerging general risks (J ablonowski, 1996).

40

2.2.4 The Goal of Risk Assessment

Specifically, risk assessment is conducted to assist in decision-making and rest of
the risk analysis process (Aven, 2003; CFSAN/F DA, 2002; Kaplan, 1997; Nader, 2002).
Once the hazard is identified, the data and insights are applied to assessing the risk,
allowing for the next step, risk management.

The risk assessment process helps to organize and quantify the hazards, to clarify
the overall risk as well as the interaction of the components comprising the total risk.
The risk assessment process has continued to evolve in complexity, but has also increased
in transparency (CF SAN/FDA, 2003; Claycarnp, 2003; Clemens & Swallom, 2005;

Office of Management and Budget, 2006; WHO, 2003).

2.2.5 Basics of Risk Assessment -— Categorizing the

Threat

Risk assessment focuses on identifying what a type of risk is, and is not, to better
evaluate the most appropriate risk model characteristics. In the most broad terms, risks
are either strategic (“long-term variations in outcomes”), operational (“traditional
business and hazard risks... variations in outcomes within one operating period”), or
economic (“political and financial situations that are created by macro- and micro
events”) (Kallman, 2007). Operational risk, most commonly defined as "the risk of loss
or harm resulting from inadequate or failed processes, people, or processes, or from an
external event" (Taylor, 2006). By these definitions, the counterfeit threat is an

operational risk.

41

To further categorize the counterfeit risk, it is valuable to review risks one step

further. Hazards can fall into several of the risk classes. This summary is based on

Nader’s “The Manager’s Concise Guide to Risk (2002)” (Nader, 2002).

Catastrophic risk affects industries and firms whose financial assets are
exposed to catastrophic natural perils, such as earthquakes, hurricanes,
volcanic eruption, and so on. . .. Although catastrophic risk is usually
considered as an outcome of natural perils, one nonnatural peril, terrorism,
has emerged in recent years as a source of risk whose consequences for
the insurance industry appear increasingly capable of attaining the same
dimensions as those of catastrophic risk.

F oreseeable risk is primarily a legal definition derived fi'om the concept of
“foreseeability.” Accordingly, foreseeable risk is any risk whose
consequences can reasonably be expected to occur, by a person of
ordinary prudence.

Fraud risk [also categorized by Nader as security risk] may be defined as
the entity’s total exposure to the probable misconduct, dishonesty, and
deceit by internal as well as external parties. . .. What is peculiar about
fraud risk is that it can never generate any returns to the party at risk.
Therefore, no business entity acting rationally should ever voluntarily bear
or expose itself to any type or amount of fiaud risk. However, we live in
an imperfect world, it is unrealistic to expect that total elimination of fiaud
risk can be achieved.

Fundamental risks are impersonal in nature, and any person affected by
such risks is exposed to losses that do not arise from that person’s own
individual choice or behavior, but from events beyond his or her control.
Examples of such events include natural disasters, political and social
developments, economy-wide phenomena, industry-wide phenomena, and
so on.

“Legal risk is one of the risks of doing international business. It arises
fi'om the weaknesses, incompleteness, nonenforceability, and other similar
problems with a foreign country’s laws and its legal-judiciary

machinery. . .. Such problems increase the probability that the legal system
will fail to provide adequate protection of physical and intellectual
property rights, or remedies against breaches of contracts and other
violations of contractual rights.

The term liability risk is applied to a very broad category of pure risks,
many of which are insurable.... Liability risk arises whenever one party
is exposed to possible loss of present or future assets or income as a result
of causing one or more of the following events to another party or to assets
owned by another party, whether those events (torts) are caused by the
first party willfully or though negligence.

The term marketing risk is used to describe the uncertainty that surrounds
the future demand for a firm’s products as a result of numerous variables

42

which affect this demand, but may be unpredictable or not entirely under
the firm’s control. Marketing Risks arise fi‘om unanticipated or
uncontrollable shift in any of the factors which affect the firm’s marketing
mix (product, price, place, and promotion).

Particular risks are those types of risk whose consequences affect
individuals separately, and are not so pervasive (as in the case of
fundamental risks) as to affect an entire group of individuals. Particular
risks arise from personal actions or events that are under an individual’s
control, and are therefore considered to be the responsibility of the
individual, rather than the responsibility of society as a whole.

Property risks encompass all events which carry a possibility of loss, to a
property owner, of one or more of the following: the value of property
(direct loss), the use of property (indirect loss), and the future income
generated by property (indirect loss).

A pure risk is defined as any risk which can only result in a loss or no loss,
but can never generate any gains to the party at risk. In other words, a
pure risk consists entirely of downside risk and does not contain any
upside risk component. . .. The designation of some risks as pure risks is
useful for setting apart those risks that are normally insurable risks fi'om
speculative risks, which are normally uninsurable risks.

Speculative risk, as distinguished from pure risk, is a term applied to
describe all risky situations that, in addition to carrying the possibility of
loss, also carry the possibility of gain to the party at risk. In other words,
speculative risks incorporate not only a downside risk component but an
upside risk component.

Depending on the specific threat and an organization’s response to risk, counterfeiting

could fall into one or many of the risk classes. As with many aspects of risk analysis,

there is often some confusion on use of terminology. Mil-Std-882D refers to an

operational risk matrix as a process within, what it refers to as, the overall operational

risk assessment program.

For this research, the FDA reference will be used.

43

2.2.6 Basics of Risk Assessment — Taking the First

Step

A common challenge faced in risk assessment is a lack of appropriate historical
data, a basic lack of knowledge important in decision-making and data that is not yet
available (Cruz, 2002; Van Der Fels-Klerx, Goossens et al., 2002). One common method
used for taking the first step is peer consultation or expert panels. Some examples are
discussed in “Building confidence in models for multiple audiences” (Howick, Eden et
al., 2008), and “Engaging Expert Peers in the Development of Risk Assessments”
(Patterson, Meek et al., 2007). A tried and true method is to apply expert judgment or the
formal Department of Defense (DOD) process of Delphi (Coffin, 2002; Cruz, 2002; Van
Der Fels-Klerx, Goossens et al., 2002). Delphi was developed by the RAND Corporation
after World War II to combine expert opinions while eliminating bias. Delphi collects
individual surveys then conducts group reviews of the overall perspective, in a very
formal and methodical process (Coffin, 2002).

The Delphi process is described here (Cruz, 2002):

Usually, one or more of the following properties of the problem to be

solved lead to the need for Delphi:

0 The problem does not lend itself to precise analytical techniques but
can benefit from subjective judgments on a collective basis

0 The individuals needed to contribute to the examination of a broad or
complex problem have no history of adequate communication and may
represent diverse background in respect of experience or expertise.
Time and cost make fi'equent group meetings infeasible
More individuals are needed than can effectively interact in a face-to-
face exchange

44

2.2.7 Basics of Risk Assessment — Appropriate

Precision

An expert or group of experts can provide quick insights, as well as identify
potential influences that may not show up until much later using data-intensive statistical
methods. A limitation of using an expert’s judgment is the role assumptions play in the
judgment. Techniques that simplify an analysis can compound into greater uncertainty in
the final output of the model (Claycamp, 2006; WHO, 2004). Usually, these
heuristically-derived assumptions (using expert judgment rather than hard, unchanging
data) are not clearly defined in the statistical review, and yet reports present the results as
very precise.

In their report “Understanding Risk”, the National Research Center (NRC) at the
National Academies of Science discussed related analysis that is conducted to reduce the
complexity of the model, and the danger that over simplifying complex,
multidimensional problems can lead to results that are “highly contentious” to the
decision-making process (National Research Council, 1996).

Additional insight was provided by Hassenzhal (2006):

Risk analysts are often asked to generate precise numerical calculations.

The expectation serves an important role in risk evaluations by forcing

discipline on the analysts. At the same time, however, the act of

generating such precise numerical calculations permits the unfortunate

possibility that the resulting estimates will be interpreted as sufficient and

accurate depictions of the risk. Alternatively, analysts may feel compelled

to present estimates that are more precise than they feel is warranted.

Stripped of quantitative and qualitative information about uncertainty,

these precise estimates may become reified and thereafter be interpreted as
the "true" or "actual" value of the risk.

45

fact. ?i
then I
OMB.
3000;}
P2000
Starr.
malts
Sung?

i C3133

1m

”ftur
c.
--:.
“I“ 51-11%.-
4"“
1
9|! ‘

The concern is that the analysis may reach a point of “excessive precision” or a
“false sense of precision and it is an important point to remember in making preliminary
risk assessments or in reviewing emerging risks (Hassenzahl, 2006; J ablonowski, 2005;

Pittinger, Brennan et al., 2003).
2.2.8 Basics of Risk Assessment — Words or Numbers

Risk assessments do not need to strive for quantitative output (numerical data). In
fact, there is a body of literature which supports qualitative (high/medium/low) output,
when there is qualitative (high/medium/low) input, including agencies such as the FDA,
OMB, DOD, international standards organizations, and industry associations (Broder,
2000; J ablonowski, 1994). For example, in a CFSAN/F DA presentation, “Tools for
Prioritizing Food Safety Concerns: An FDA Perspective,” (by Dr. Robert Buchanan,
Senior Science adviser and Director of Science) both qualitative and quantitative tools or
models for risk ranking or assessment were validated, noting that each can have their
strengths and drawbacks (Buchanan, 2007). This is supported by other authors, as well
(Claycamp & Hooberman, 2004; Etherton, Main et al., 2008; Shepherd, Barker et al.,
2006).

In a study in Risk Management by Jablonowski (1994), a survey of risk managers
found that words communicated the risks better than numbers. For example, it was more
descriptive and helpful to call something “rare” than to refer to a “0.05” percent chance
of occurrence.

The lack of detailed history is also echoed by the security industry. For example,

the book “Risk Analysis and the Security Survey” (which is recommended reading for

the American Society for Industrial Security ASIS Certified Protection Professional

46

certification) specifically addresses security-related risks that apply directly to the anti-
counterfeit strategy situation (Broder, 2000):

Threat occurrence rates and probabilities are best developed from reports

of occurrences or incident reports, whenever this historical data exist.

Where information does not exist, it may be necessary to reconstruct it.

This can be accomplished by conducting interviews with knowledgeable

persons or by projecting data based upon educate guesses, supported by
studies in like industries and locations.

Review of the literature and other reports did not specify how to separate the
categories (hi gh/medium/low) beyond stating that the categories should be spaced far
enough apart to be meaningful. The categories should not be set to parts-per-trillion if
the available equipment can only measure to part-per-million (the data would be
impractical to measure). Likewise, the categories should not be set so all the responses

fall into the same rank (if all are “medium” there is no differentiation).

2.2.9 Basics of Risk Assessment — Numerical Methods

This section is based on the book Numerical Methods for Engineer” by Steven
Capra and Raymond Canale (1998). Traditional numerical models, formulated to be
solved with arithmetic operations, are the focus of engineering textbooks (Capra &
Canale, 1998). Unlike engineering problems (e.g. will a bridge withstand a certain load),
the counterfeit risk has human factors and there may be no physical laws in the root
cause. Due to the nature of the data, even traditional “fundamental laws” (e. g. energy
cannot be created or destroyed) cannot be derived fiom empirical tests (e. g. observation
or experiment) (Capra & Canale, 1998). The Counterfeit Product Risk Model is not a

physical system or process that is easily defined for future threats.

47

Since engineering problem-solving concepts are familiar to many of the engineers

assigned with counterfeit risk prediction, these concepts will be reviewed for their

relevance to the model:

Flowcharting. While flowcharting of the processes can work in a very general sense
for counterfeit risk, the lack of large, historical data sets reduces the effectiveness of
developing numerical models.

Analytical versus numerical methods. Analytical methods are considered exact,
approximated by linear functions, and are of limited value since most “real” problems
are more complex (Capra & Canale, 1998). Numerical methods are based on
“equations, non-linearity, and complicated geometries” and are very common in
engineering problem solving (Capra & Canale, 1998). Even though the numerical
solutions can be quite complex, they are still an approximation. “Although perfection
is a laudable goal, it is rarely, if ever, attained” (Capra & Canale, 1998). Neither
applies to the Counterfeit Product Risk Model due to the underlying lack of historical
data and the ever-changing nature of the risk.

Truncation versus round-off errors. Truncation error is a different concept that
measures the variation in the approximation of a number and the actual number (e. g. a
measured number 4.859 but truncated to 4.9) (Capra & Canale, 1998). Round-of
error is due to method or computers simplifying the data to a finite number (e. g. Pi is
3.141592653589... but rounded off to 3.14) (Capra & Canale, 1998). These concepts
both assume there are a large historical data sets and some accurate measure of the

system. The only historical data about counterfeiting is whether the product is known

48

to be counterfeited and how many occurrences of the fake product were identified.
Thus, neither concept applies to the Counterfeit Product Risk Model.

Significant figures, accuracy, and precision. Significant figures refer to the
number of digits that can be used to represent the data point (Capra & Canale, 1998).
Accuracy is how closely a measured result equates to the actual result, whereas
inaccuracy is the deviation from the actual result (e. g. inaccurate data points are far
from the actual number) (Capra & Canale, 1998). Precision refers to how two
computed data points agree with each other regardless of how close they are to the
actual number, whereas imprecision (or uncertainty or bias) is the magnitude of the
variation in the computed data points (e. g. precise results can be data points that are
close together regardless of the location of the actual number) (Capra & Canale,
1998). Since the actual number for a counterfeit risk calculation cannot be known,
these concepts do not apply to the Counterfeit Product Risk Model.

“Blunders,” formulation errors, and data uncertainty. Engineering problem-
solving assumes that the model and assumptions are sound, but in the real world that
should not be assumed. “Blunders ” are considered to be gross errors in the data
gathered, or models that are not fundamentally sound relative to the data collection or
measurement (Capra & Canale, 1998). Formulation errors, or model errors, are fi'om
“incomplete mathematical models” (Capra & Canale, 1998). If a counterfeit risk
model is based on the analysis of past data, then a mathematically representative
model can be developed, but in trying to predict future risks the model developed
would be fundamentally incomplete. With these fundamentally incomplete models,

uncertainty (bias) would be associated with the results. When the counterfeit risk is

49

considering future risk, the uncertainty could magnify, detracting from the value of
the prediction. Data uncertainty is the error from the physical data that was used to
build the model (Capra & Canale, 1998). The physical data used to build a numerical
counterfeit model would be both uncertain and imprecise.

Quantitative methods are not always the most efficient or most appropriate
prediction models. Also, classical statistical methods are based on the fundamental
concepts outlined above, and are not practical to apply to the Counterfeit Product Risk
Model. The use of classical statistical tools such as mean, standard deviation, or variance
would give an impression of excessive precision.

If the Counterfeit Product Risk Model used a traditional analytical or numerical
model, the associated error would lead the output to be impractical in practice. This

review supports using a more general, qualitative model.

2.2.10 Basics of Risk Assessment—Risk Presentation

This section will review potential risk presentations which may have application
to the Counterfeit Product Risk Model. An investigation of the factors that make up the
probability categories, in the probability versus severity matrix has been conducted using
the literature (Table 4). The operational risk matrix is an application of the basic
probability versus severity matrix. Operational risk management (ORM) in the form of a
matrix is frequently referred to by the World Customs Organization (W CO), FDA, and
others (Acheson, 2007b; Brackett, 2004; Elsworth, 2007; Mil-Std-882D, 2000; World

Customs Organization, 2003).

50

Table 4. A General Probability versus Severity Matrix

 

 

 

 

 

 

 

 

 

 

SEVERITY

Catastrophic Critical Marginal Negligible
E Frequent 1 3 7 l3
:1 Probable 2 5 9 16
fl Occasional 4 6 l 1 18
8 Remote 8 10 14 19
E Improbable 12 15 17 20

 

 

 

 

 

 

 

When considering food defense risk, David Acheson, the Assistant Commissioner
for Food Protection for the US FDA, identified several key concepts in a presentation he
gave (Acheson, 2007b):

0 Vulnerability Assessment Tools
0 Operational Risk Management (ORM)
o Carver plus Shock
I Federal only
' SPPA
I Software tool
0 Strategic Partnership Program on Agroterrorism (SPPA)
o Raise Awareness
- ALERT

' Web-based training
The Food and Drug Administration’s ALERT (Assure, Look, Employees,
Reports, and Threat) initiative is specifically designed to raise awareness by fi'ont line
employees in retail and operations, of actions and anomalies that could constitute
dangerous food defense risks that would cause harm (CFSAN/F DA, 2006).

The ORM process is used by a wide range of groups, and several are referenced in

Clemens & Swallom (2005) including:

51

Naval Ordnance Safety and Security Activity,

US Federal Aviation Administration (FAA),

US. Army Aviation and Missile Command,

US. Army Research Development and Engineering Command’s Aviation
Engineering Directorate,

Swedish Defense Materiel Administration (Forsvarets Materielverk),

US. Army Apache Attack Helicopter Project Management Office (PMO), and
the Office of the Director of Army Safety.

The Operational Risk Matrix was central to all the above examples, as were the
qualitative factors associated with the most basic probability and severity categories:
Critical, Marginal, Negligible, Frequent, Probable, Occasional, Remote, Improbable,
Extremely Improbable (Clemens & Swallom, 2005; Mil-Std-882D, 2000). The
quantitative range for the categories was determined by experts who interpreted the data.
In some cases, such as aircraft crashes, exact data was known and could be directly
applied (Clemens & Swallom, 2005).

The Clemens article clearly identifies the underlying assumption that the
probability and severity category rankings are variable and based on interpretation of
unknown data. For example, looking at several of the nine probability categories, it is
demonstrated that the input data relies on science and expert opinion: Frequent - once
every 2 days, Unlikely - once every 10 years, and Extremely Improbable - once in 1000
yrs. The final assumption is quantitative, but the underlying selection of the range of
risks is still heuristic or biased. There should be a definition of the range of each
category, for example, why isn’t “frequent” considered to be “once every two weeks”?

The risk analysis process for a specific risk or event should consider how the
information will be used and how the information must integrate into over-arching risk
management systems (Acheson, 2007b; Buchanan, 2007; Clemens & Swallom, 2005). A

common practice is “risk summing,” also referred to as using a “risk register” or Total

52

Risk Profiling (TRP), where risks across a system or organization are gathered and ‘
consolidated in an overall review (Clemens & Swallom, 2005).

The concept of risk is growing in importance for agencies and industry for a
number of reasons, including the increased cost and legal liability of recalls and the US
Securities and Exchange Commission (SEC) reporting requirements of the Sarbanes-
Oxley Act (DHS, 2007). An emerging trend is for companies to use an Enterprise Risk
Management Concept (ERM) and to assign a Chief Risk Officer (CRO) who would
report directly to the Chief Executive Officer (CEO) of a corporation (Lam, 2003; Lee &
Shimpi, 2005; Phillips, 2008; United States Government Accountability Office, 2008b).

The Clemmons (2005) report summarizes the risk-summing requirements
(Clemens & Swallom, 2005). A risk summation should:

Characterize total system risk

Adapt to both quantitative and qualitative methods

Produce a unique, unambiguous result

Be easy to calculate

Be easy to interpret

Serve risk communication needs well

Satisfy needs for transparency, clarity, consistency and reasonableness
Provide a clear path to the required decision

For the Counterfeit Product Risk Model, the rather simple nature of the
operational risk management and operational risk matrix are appropriate for the very
small amounts of historical data. Specifically, the OMB states that “when large amounts
of historic data from humans are available, an actuarial risk assessment may be
performed using classical statistical tools” (Office of Management and Budget, 2006).
The classical statistical tools that do not apply to the broad and ambiguous counterfeit
risk concept include decision trees, Monte Carlo simulation, probabilistic risk analysis

(PRA), Bayesian analysis, etc. The classical statistical tools do apply when there is more

53

information or a more defined process. For example, the World Customs Organization
(W CO) uses classical statistical methods to review their actual inspection data (risk
assessment), but uses insights on what to actually inspect in the first place (hazard
identification) (World Customs Organization, 2006). l

The operational risk management (ORM) has merit and can be used for the

Counterfeit Product Risk Model assessment.

2.2.11 Basics of Risk Assessment—Model Building

When reviewing risk model building, it was clear that there is no one best model
(Acheson, 2007b; Claycarnp, 2003; Cruz, 2002; Hassenzahl, 2006; Ryan, 2005). Even
the formal standards organizations (ANSI, ISO, EN, Mil-Std, etc.) only begin with a risk
assessment framework and do not recommend a best model (Etherton, 2007). Often, as
recommended above, the model selected is similar to other models used in that field or
area: food safety, corporate security, law, finance, insurance, operations, etc.

Models to assess the counterfeit threat are, thus, specific to the industry, company
or Agency. To harmonize the information, companies and agencies measure the
counterfeit risks and factors in terms of their other models which deal with risks and
threats. An inherent limitation of these models is that they may not consider a broad
enough view of the counterfeit threat to really expose the root causes of the risk or threat.
In addition, these models may not be efficient or effective to evaluate the effectiveness of
potential countermeasures or deterrents.

Specific standards were analyzed and summarized, to determine if they have
application to the anti-counterfeit prediction of risk. Generally, each presents a risk

framework but few give detail on conducting the actual risk assessment.

54

EN 1050. The European Union (EU) developed this “new approach to technical
harmonization and standards) in the 1990’s, and it includes a specific machinery
safety focus” (Etherton, 2007). The standard includes guidance on machinery design
and includes EN1050, the “Safety of Machinery: Risk Assessment (Etherton, 2007).
ANSI B11 TR3. The American National Standards Institute (ANSI) Machinery Tool
Safety Standards Committee formed the TR3 guidelines in 1995. This standard
raised the US standards up to, or above, the EN 1050. The general steps for TR3
machine risk assessment are: (1) identify the tasks and hazards; (2) estimate the risks;
(3) determine the steps necessary for reducing risk; (4) select safeguards; (5) verify
whether risks were reduced; and (6) document the reduction of risks. The risk of a
task is determined by evaluating the relationship between the potential probability of
an occurrence of harm (e.g., very likely, likely, unlikely, and remote) and the
potential severity of harm (e.g., catastrophic, serious, moderate, or minor). These
factors are entered into a risk estimation matrix or equivalent risk estimation tool, and
a risk level (e.g., high, medium, low, negligible) is then determined (Etherton, 2007).
ISO 14121. The International Standards Organization (ISO) formed the TC199
“Safety of Machinery” technical committee to focus on machinery safety. The ISO
work group developed this standard.

MIL-STD-882D. The DoD developed and uses “one of the most pervasive risk
assessment approaches” in this “Standard Practice for System Safety,” which is used
by “all departments and agencies within DoD.” “As with all existing safety

programs, the objective of this document is to develop a standardized method of

55

 

assessing and mitigating risks” (Etherton, 2007). The standard states that it complies
with all other US Government standards.

Since it has a direct application to the anti-counterfeit prediction of risk, Mil-Std-
882D will be reviewed in more detail (Mil-Std-882D, 2000). The objective of this
Standard is to develop a standardized method of assessing and mitigating risks. The
Standard is designed to protect private and public personnel from “accidental death,
injury, or occupational illness; weapon systems, equipment, material, and facilities from
accidental destruction or damage; and public property while executing its mission of
national defense.” The Standard states that it complies with all other US Government
standards. The Standard refers to events as “mishaps” and the risk as “mishap risk.” The
Standard does refer to mishaps in DoD operations, so the term may be, essentially, a
subcategory of overall operational risks.

As was discussed in the previous risk assessment discussions (and this is a
common challenge for developing a Counterfeit Product Risk Model) the process for
identifying hazards is not specified but there is an emphasis on “historical hazard and
mishap da ” and on reviewing “lessons learned from other systems.” The process to
define the hazards or hazard categories are not identified in the Standard, though it is
specified that the mishap risk assessment values are grouped into mishap risk categories.

The Standard is actualized using a “mishap risk assessment matrix.” This matrix
is known by many terms, such as an Operations risk assessment or a “probability x
severity” matrix, but the basic components are a quantitative probability measure on one

axis and a quantitative severity measure on the opposite axis. Utilizing the matrix allows

56

the risk manager to assign a “value to a hazard based on its mishap severity and its

mishap probability.”

2.2.12 Practical Application of Risk Assessment

This practical example is included in this Risk Overview Chapter, and not the
applied Risk Model Chapter, since it demonstrates the risk analysis concepts. Richard
Ryan is a Corporate Security manager at Archer Daniels Midland (ADM) Company and
he presented “Risk Assessment to Drive Research for Contaminant Detection” at the
Institute of Food Technologist’s 1st Annual Food Protection and Defense Research
Conference on November 4, 2005 (Ryan, 2005). This presentation was especially
valuable to the Counterfeit Product Risk Model research because it emphasizes real
actions and concerns at a food company. The presentation focused on all intentional
contamination or attack risks. The following is based on Ryan’s presentation.

“Security upgrades” are operational costs that do not create revenue, but they do
reduce costs or liability, and the cost and benefit analysis is presented as a business case.
When analyzing a risk mitigation project, three keys were discussed. The first key is the
risk assessment. Although ADM uses the Carver+Shock model as the base of their risk
assessment program, they clearly emphasize that the model does not consider every risk,
so the results are admittedly incomplete. “[Carver+Shock] works perfectly well if you
know who the bad guy is and what his motivations are. Unfortunately, making an
assumption as to the identity and motivation of the attacker introduces unavoidable error
in the risk assessment of this type” (Ryan, 2005).

With usable but incomplete models in use, it is not appropriate to affix specific

actuarial numbers to the risk (“risk rating is 3”). In industry and government there must

57

be some more subjective analysis to reflect the risk. The confidence in the probability is
stronger if there is reference to other scientific research or models regarding the risk, to
help change the perception fiom a perceived risk to a real risk. The confidence also
increases with anecdotal evidence or details of counterfeiting in other industries or
product groups. In the subjective analysis included in the business case, a range of
potential financial liability would be determined.

The second key is to determine the potential mitigation steps and to choose the
most efficient and effective solution — lowest cost and highest impact. And finally, the
third step looks at what level of confidence there is that the mitigation step will be

effective.

2.2.13 Risk Concept Summary

This section reviewed risk analysis, and its three components (risk assessment,
risk management, and risk communication), to focus on what does, and does not, apply to
the Counterfeit Product Risk Model. The science of risk assessment has become very
analytical and quantitative, but the nature of emerging risks, that do not have large
historical data pools, supports a more manageable and more qualitative
(high/medium/low), heuristic approach to risk management and risk assessment. This
section also reviewed how the incomplete nature of the data and the clandestine nature of
the counterfeit threat lead to inefficiency when trying to use classical statistical methods,
such as the Monte Carlo simulation, decision trees, or probabilistic risk assessment
(PRA). There is wide acceptance of quantitative, semi-quantitative, and qualitative risk
management and risk assessment tools. The literature supports using an expert panel to

develop the Counterfeit Product Risk Model. More generally, the models and tools

58

should be able to be used together; there is acceptance of building unique risk models to
support a unique risk.

The probability versus severity matrix, used by operational risk management, is
the most efficient and effective type of structure for this research. The literature and
agency reports support dissecting the overall probability into smaller and more detailed
factors. This discussion will continue in the next chapter, with more specific examples.
The risk analysis and risk assessment concepts and tools have been used in developing

the Counterfeit Product Risk Model.

59

2.3 Risk Model Research

This chapter will review risk analysis models and processes to focus on what can,
and cannot, apply to the Counterfeit Product Risk Model, using the literature and agency
reports. There have been many models inefficiently applied, and others that are assumed
to be applicable, to the counterfeit threat. Thus, it is important to review model theory as
specific models to determine if they can be used to assess the counterfeit threat.

The lack of refereed articles, and credible counterfeit data, is mirrored by a lack of
risk assessment methodology and models relevant to counterfeiting. The lack of even the
most basic quantitative factors is due to the nature of the threat—being clandestine,
criminal, and rapidly evolving in complexity and range. The counterfeiters have
clandestine operations and in many cases the product is so good that it travels through the
supply chain undetected.

In general, information on the gaps or breaches is considered classified by
agencies or confidential by industry. Any publication of these gaps and breaches could
be read and used by the “fraudsters”l to help them be more efficient in their operations.
The lack of public information or data leads to a very limited historical resource.
Classical, statistical, data-intensive risk assessment models or tools are not sufficiently

useful.

 

' The term “Fraudster” is used as a descriptive, formal term for this specific type of
criminal and their activity (deKieffer, 2006b; European Commission, 2006;
PriceWaterhouseCooper, 2007).

60

 

2.3.1 Risk Assessment Models Overview

This Section reviews specific risk assessment tools from a wide range of
applications to assist in the selection of the most appropriate Counterfeit Product Risk
Model. Risk assessment is included in the standards of several key intemationally—
recognized standards setting organizations. Each of the standards uses the simple,
traditional “probability versus severity” matrix in an operational risk management (ORM)
assessment. ORM is recommended by, and used by the FDA, DOD, USDA, and others.
ORM is also the base concept used in risk analysis, notably in Mil-Std, ANSI, ISO, and
EN standards.

In these standards, the selection of quantitative probability and severity data are
left up to the risk managers, though in many cases, an additional expert panel is
recommended. For example, the ANSI B11 TR3 emphasizes risk analysis processes that
are commonly referred to in other processes (Etherton, 2007). These include:

identify the tasks and hazards;

estimate the risks;

determine the steps necessary for reducing risk;
select safeguards;

verify whether risks were reduced; and
document the reduction of risks.

Similarly, there are risk categories and identifiers (Etherton, 2007): g

o evaluate severity versus probability
0 common severity categories - catastrophic, serious, moderate, or minor
0 common probability - very likely, likely, unlikely, and remote

Depending on the model or standard, a number of categories can be defined to
meet the needs of the data. The standards use anywhere fiom three to nine categories of

inputs.

61

2.3.2 Risk Assessment Models in the Anti-Counterfeit

Industry

Any precise risk assessment will include a perspective on very sensitive gaps or
weaknesses in the supply chain, which, if published, would provide insight to criminals
that would put that supply chain at even greater risk. The Department of Homeland
Security has classified reports or reviews that divulge too much information about gaps
or that create a “roadmap” which could lead to a breach of security (CFSAN/F DA,
2007c; Congressional Research Service, 2003; Department of Homeland Security, 2007;
Scandia National Laboratory, 2007; Tilden, 2007). The FDA procedures recommend that
companies hold their reports confidential and DHS has stated that their detailed reports
have been classified as secret (CFSAN/FDA, 2007c, 2007d, 2007c; Collins, 2004;
Department of Homeland Security, 2007; Scandia National Laboratory, 2007). Although
not often enacted, “[information] classification on national security grounds denies access
to anyone without a govemment-issued security clearance and a demonstrated ‘need-to-
know’ the information” (Center for Strategic and International Studies, 2005).

Counterfeit risk assessment based on expert panel or heuristic insights are being
conducted, to some extent, but details are not available in the public realm, and thus, are
considered confidential (Hammerbeck, 2008; Pierce, 2004). To show how little is
published in regards to counterfeit risk assessment is to review the 2004 final report and
the June 2006 update of the FDA report on “Combating Counterfeit Drugs” (FDA, 2004,
2006a). This series of reports and updates only mention a high level FDA intent to

“support the development” of several of the counterfeit risk predictors.

62

The FDA has taken an initial step by noting some potential categories that may
predict risk. Past counterfeiting is a predictor of future counterfeiting and that diversion
is a gateway for illegal products to enter the supply chain (deKieffer, 2006b; Liang,
2006). The Carver+Shock risk assessment tool does contain one question regarding
whether the manufacturing process includes known counterfeit or diverted products
(CFSAN/F DA, 2007a). A review of the FDA Combating Counterfeit Drugs report in
2004, and the 2006 Update, and the full-text is included here (FDA, 2004, 2006a):

There was agreement that the criteria we suggested to identify drugs that
were likely to be counterfeited were correct. These included:

Impact on public health if the drug were counterfeited;
0 Drugs history of counterfeiting;

0 Drugs price;

0 Drugs volume;

0 Drugs dosage form;

0 Drugs clinical uses; and

0 Whether similar products had a history of being counterfeited.

However, there was no consensus on how to apply these, or other, criteria
in creating a list of such products.

As stated above, some comments suggested that instead of developing a
list of drugs likely to be counterfeited, a set of criteria for determining
whether a drug was [as] likely to be counterfeited should be created. One
proposal for such criteria was:
0 A drug has been subjected to a seizure or stop sale notice because
of counterfeiting, or
0 There is documentation that a drug was counterfeited and is the
subject of an investigation by federal or state authorities, AND
0 The product is high cost (e.g., over $200 per dose) or high volume
(e.g., top fifty drugs), or
o The product is used extensively for treatment of HIV/AIDS or
cancer, or
The product is injectable, or
The product distributed in a special or limited way, or
There are multiple documented instances of pedigrees not being
passed with the product

63

Ir]

There are some models that review the counterfeit threat, but none were found

that support a comprehensive Counterfeit Product Risk Model.

2.3.3 Model Example - Theoretical Smuggling Model

Ivanova (2007) uses a series of three predictions of how their formula-based
model “Corruption, illegal trade and compliance with the Montreal Protocol” would react
to several situations. Specifically, Ivanova studied the impact of only three factors
(corruption, rule of law, and tariffs) on chlorofluorocarbons (CFC) smuggling. The
author does state that this is a partial review of the issue. The author tests the model
against the predictions (whether the model points in the expected direction).

Ivanova supports that idea that testing a comprehensive model using a classical
statistical method to predict smuggling (the data has similar specifics as counterfeiting) is
not possible. “Since data on illegal trade is not available, it is not possible to test any
theoretical conclusions about smuggling” (Ivanova, 2007).

Neither the FDA report, nor any other agencies found in this research,
recommended a specific risk assessment methodology nor did they provide any specific

details of previous risk assessment projects.

2.3.4 Model Example: GTRIC

The General Trade-Related Index of Counterfeiting (GTRIC) model is presented
in the, Organization for Economic Cooperation Development (OECD) report, titled
Economic Impacts of Counterfeiting and Piracy (OECD, 2007a). This report was

published as a draft report in September 2007.

64

The OECD was originally formed in 1947 to administer the Marshal Plan for

European reconstruction after World War H and has had a continually evolving mission

to help member countries achieve economic growth. There are thirty members (including

the G8 and others), and over 70 participating developing and emerging market

economies.

The OECD’s Economic Impacts of Counterfeiting and Piracy report is the most

complete and recent agency overview on counterfeiting. The contributors or editors are

not identified in the draft report.

The report emphasizes the difficulty in product counterfeiting data collection:

The overall degree to which products are being counterfeited and pirated
is unknown and there do not appear to be any methodologies which could
be employed to develop an acceptable overall estimate (OECD, 2007c).

The response rate to the [OECD] Customs questionnaire was not high
from larger economies, which limited the usefirlness of the results. As for
the data itself, sizeable fluctuations fi'om year to year in many economies
made it difficult to observe any trends or patterns; moreover, in many
cases the level of inceptions/ seizures reported, when compared to other
economies, raised questions about the consistency and comparability of
the information (OECD, 2007c).

With regard to comparability, it should be noted that the basis for
reporting interception values varied among respondents. Some reported
on the basis of declared customs value while others reported market value
or legitimate item value (OECD, 2007c).

Carrying out assessments of the factors (or drivers), even on a qualitative,
nonempirical basis, can generate insights into the counterfeiting and piracy
situation in different products and in different economies (OECD, 2007c).

Regardless of the admitted data limitations, the report does present new

mathematical models for estimating the magnitude of the counterfeit threat. The models

estimate counterfeiting magnitude by global production (“GTRIC-p”) and source

economy (“GTRIC-e”), which is the General Trade-Related Index of Counterfeiting

(GTRIC) model. The GTRIC identifies a “ceiling” for international counterfeiting,

65

 

 

essentially establishing an “upper fix-point limit” (OECD, 2007c). The data is generated
in three steps: (1) identify known counterfeited product, (2) identify known counterfeit
producing source economies, and (3) estimate the rate of counterfeiting based on customs
seizure and OECD surveys. The GTRIC model did not include any direct references to
authors or sources and there were no additional literature references on the topic.

The model states an impressive 95% confidence interval, but the prediction range
is extremely broad. For example, the GTRIC-e estimate for Morocco, with a 95%
confidence interval, is that between 10% and 97% of the exported products are
counterfeit.

The OECD Economic Impacts of Counterfeiting and Piracy report examines the
drivers and root causes of counterfeiting from a global, country, or general-industry
perspective. The OECD report has many general insights that can be applied to product
or company specific assessments, but the GTRIC models do not contribute to the
Counterfeit Risk Model because they focus on generalizations about countries not a

specific company’s product.
2.3.5 Risk Assessment Models in the Food Industry

There are a number of overall risk analysis and specific risk assessment models
used in the food industry, though most refer to basic operational risk assessment when
reviewing emerging threats. Unlike counterfeits, the general food industry risk analysis
models utilize a large sample size and are very data driven (Buchanan, Dennis et al.,
2004; Buchanan, 2004). The risk analysis procedures generally focus on either pure food

safety (unintentional contamination such as E. coli or contaminants) or food defense

66

(intentional contamination with malicious intent, such as terrorist attacks, disgruntled
employee contamination, or malicious tampering).
2.3.5.1 Good Manufacturing Practices (GMP)

The Good Manufacturing Practices (GMP) procedures regulated by the US.
Government are very rigorous, driven by large data sets, and work with predetermined
hazards. The US Federal Food, Drug, and Cosmetic Act regulates that human foods must
be free from adulteration, which it defines as harmful impurities (CFSAN/F DA, 2001b).

The Code of Federal Regulations (CFR), GMP covers four main areas:

a 21 CFR Parts 210 and 211 — Drugs
0 21 CFR Part 820 - Medical Devices
0 21 CFR Part 110 — Food

0 21 CFR Part 606 - Blood

GMP does not specifically consider economic fraud, counterfeit, or risks outside
traditional operations.
2.3.5.2 Good Agricultural Practices (GAPS)

The Good Agricultural Practices (GAPS) are global programs that are procedures,
but not regulated (not law), in the United States (FDA, 1999; Food and Agriculture
Organization of the United Nations, 2002). The GAPS focus on agricultural practices,
such as farming of fresh produce, and include recommendations on traceback, packaging
and filling sanitation, and cover other processing steps (CFSAN/F DA, 2001b; Food and
Agriculture Organization of the United Nations, 2002; Smith, 1999). While the GAPS are
outlined in great detail, they do not address or add to the Counterfeit Product Risk Model

other than in their general concepts.

67

2.3.5.3 Hazard Analysis and Critical Control Point (HACCP)

The Hazard Analysis and Critical Control Point (HACCP) procedure is widely
adopted (CFSAN/FDA, 2001a). The HACCP program is covered under 21 CFR 120 (21
CFR 120, 2003; CF SAN/F DA, 2001a). The HACCP process includes a hazard analysis
of critical control points that is based on seven principles (CF SAN/F DA, 2001a).

Analyze hazards

Identify critical control points

Establish preventive measures with critical limits for each control
point

Establish procedures to monitor the critical control points

Establish corrective actions to be taken when monitoring shows that a
critical limit has not been met

Establish procedures to verify that the system is working properly
Establish effective recordkeeping to document the HACCP system

A broad interpretation of the key points, at least as applied to the shipper or the
retailer, shows that counterfeit food products could be included under the HACCP

regulation (21 CFR 120, 2003).

The written hazard analysis shall consist of at least the following:

1) Identification of food hazards;

2) An evaluation of each food hazard identified to determine if the
hazard is reasonably likely to occur and thus, constitutes a food
hazard that must be addressed in the HACCP plan. This
evaluation shall include an assessment of the severity of the illness
or injury if the food hazard occurs;. ..

The regulations do not specifically outline an emerging risk assessment model
and do not give guidance as to when a “hazard” should be included in a HACCP program
(CFSAN/FDA, 2001a). Neither do the HACCP procedures address economic fraud or

counterfeit products (Cook & Thurber, 2006).

68

2.3.5.4 Carver+Shock Risk Assessment

The Carver+Shock (called “carver-shock” or “carver plus shock”) risk assessment
method is used for reviewing intentional adulteration attack threats and for prioritizing
responses that has been applied to the food sector by DHS and the FDA. This tool is a
cornerstone of the DHS Food and Agriculture National Infrastructure Protection Plan
(N IPP) (Department of Homeland Security, 2007). Specifically, the CARVER part of the
process is a military tool developed for warfare that is now applied to terrorist threats,
who attempt to shock society by the horror of their actions, thus the “+Shock” suffix
(CFSAN/F DA, 2007a; Scandia National Laboratory, 2007).

C.A.R.V.E.R. stands for: criticality, accessibility, recuperability, vulnerability,
effect, and recognizability (Savada & Metz, 1999). “This is the only quantifiable tool I
know of for the food industry,” Frank Busta, director of the National Center for Food
Protection and Defense (Scandia National Laboratory, 2007).

Scandia laboratories has developed a Carver+Shock software tool that is
available on the FDA website. The tool surveys the user, and develops a detailed flow
chart of the operation. Each variable is given a score, and the final score is based on the
C,A,R,V,E,R, and S variables

As with other methods, such as Mil-Std-882D, applying the Carver+Shock scales
directly would bias the counterfeit product threat as a low priority. For example, the
dollar losses equated with counterfeiting would typically fall in the “low” category in the
Carver+Shock tool.. The “+Shock” considers health, psychological, and collateral

national economic impacts” (CFSAN/F DA, 2007b). The lack of an intent to shock on the

69

 

part of the counterfeiters renders the “+Shock” aspect of the tool inapplicable to the

economic fi'aud aspects of counterfeiting (see Table 5).

Table 5. Carver+Shock Software Tool: Shock Scale Descriptions

 

Shock Scale (1-10)

 

Target has little historical, cultural, religious, or other symbolic importance. Loss of life 3-4
less than 100. Small impact on sensitive subpopulations, e.g., children or elderly.
National economic impact between $100 million and $1 billion.

 

 

Target has no historical, cultural, religious, or other symbolic importance. Loss of life 1-2
less than 10. No impact on sensitive subpopulations, e.g., children or elderly. National
economic impact less than $100 million.

 

 

For example, if the “+Shock” is used to review the Tylenol poisoning incident
(the tampering of Tylenol with cyanide killed seven people and the costs were estimated
at just below $1B (GlobalOptions Incorporated, 2003)), and it would score a 1-2 on the
Carver+Shock tool regarding deaths and 3-4 regarding costs (Mitchell, 1989).
Counterfeit product incidents—the melamine in pet food, the counterfeit branded
toothpaste in dollar stores, and counterfeit Viagra—would all score a “+Shock” value of
1 -2.

The process of scrutinizing each process step for vulnerability has application in
evaluating the counterfeit threat, but since the tool is focused on attacks that have an
intent to harm, and the focus in on national-level impacts, the Carver+Shock tool does not
directly apply to pure economic fraud or to the counterfeit threat (CFSAN/F DA, 2007b;

US Department of Agriculture, 2005).

2.3.5.5 Other Food-Related Terrorism Risk Models

Many models have been developed for reviewing terrorist threats to the food
industry in terms of deaths, dollars, and other attributes, or models that try to estimate the

effectiveness of deterrent measures. Some of the models used are extensive and are

70

 

based on very complex mathematical equations to predict the outcomes (Bogen & Jones,
2006; Coffin, 2002; DHS, 2007; Ericson, 2006; Kosal, 2006; Willis & LaTourrette,
2008). These models are used to develop very broad estimates for entire industries or
countries, not for specific products or companies.

Another limitation, noted by several authors, is that the models include great
uncertainty or variability, since they use many assumptions; are only valid against known
scenarios; and apply to very general (non-specific) targets. Nevertheless, these
quantitative models are necessary in evaluating programs, particularly when legislation or
regulation is being considered.
2.3.5.6 FSIS/USDA Risk Analysis

In this section, the food related risk analysis reports and models used by the
United States Food Safety Inspection Service (F SIS) of the United States Department of
Agriculture (USDA) will be reviewed to identify what is and is not applicable to
economic fraud and the counterfeit threat. FSIS specifically covers the safety of meat,
poultry, and egg products (Federal Bureau of Investigation, 2006). FSIS has a formal
Risk Analysis program that focuses on the three traditional areas: Risk Assessment, Risk
Management, and Risk Communication (F SIS/U SDA, 2003b). A key to the Counterfeit
Product Risk Model is that FSIS hazard identification is assumed to already be completed
and qualified. The assessments begin with a general operational risk management
(ORM).

FSIS states that risk assessment may be qualitative, semi-qualitative, or
quantitative. The report also notes that qualitative assessments are used to review

generalized risks and that they result in levels of risk such as high, medium, or low.

71

”if?

eighm
@1101}; )‘c [530
F0110 R ,3
simmmm
“Whit and
W10130103:

[kink

Semi-quantitative assessments are used to “prioritize risks in relation to one another,”
while quantitative assessments are used to identify and evaluate control points or the
effectiveness of intervention strategies.

Generally, the FSIS risk assessment begins with a step that clarifies the subject of
the assessment, then evaluates adverse health effects, reviews the severity, and finally,
constructs a dose response model (to determine the extent of the risk per dose of the
hazard).

Several examples of past FSIS risk assessments are: Salmonella enteritidis in
eggs, E. coli 0157:H7 in ground beef, and Listeria monocytogenes in ready-to-eat meat
and poultry products. Each of these published, formal risk assessments contained large
data sets that were gathered in a laboratory setting and then had advanced statistical
analysis applied to them (FSIS/USDA, 1998, 2003a, 2005). The fact that these reports
were not classified, and other, detailed Department of Homeland Security National
Infi’astructure Protection Plan (DHS NIPP) reports were, confirms the sensitivity of the
economic fraud and food defense assessments.

Following Risk Analysis, there is a review of Risk Management which involves
evaluating policy decisions, and Risk Communication, which involves communicating to
the public and other risk assessors. There are no published FSIS risk assessment

methodologies that apply directly to the Counterfeit Product Risk Model.
2.3.5.7 CFSAN/FDA Risk Analysis

The Center for Food Safety and Nutrition (CF SAN, referred to as “siff-san”) of
the United States Food and Drug Administration (FDA) has risk analysis and risk

assessment processes and procedures that are aligned with the USDA and with other US

72

docufi

 

 

 

and international standards. The CF SAN overview was discussed in detail earlier in this
document in explaining the basics of risk assessment, which included HACCP and GMP.
As with the F SIS, most of the published risk assessment reports have very large
data sets and contain predetermined hazard identifications. Hazard identification is
conducted with a science- and risk-based approach, but it is usually done within groups

that have security clearances (Spink, 2008).

2.3.6 Other Specific Examples

2.3.6.1 Example - Zurich Hazard Analysis (ZHA)

The Zurich Hazard Analysis (ZHA) is a version of the traditional operational risk
assessment (probability x severity matrix) used by the Zurich Insurance Group for the last
20 years. It is a proprietary application of the general risk assessment concepts:
identification, assessment, evaluation, reduction measures, and a review of residual risk.
Residual risk is the risk that remains after risk mitigation actions are taken. The ZHA
includes a step for boundary definition of the total level of acceptable risk. The matrix
then defines specific boundaries for acceptable/unacceptable risk levels (see Table 6).
The insurance company and the client businesses determine the boundary conditions of
acceptable and unacceptable.

Mark Moorman, Vice President of Food Safety for the Kellogg Company,
discussed the ZHA concept during a Seminar at the National Food Safety and Toxicology

Center (NFSTC) at Michigan State University on June 13, 2007 (Moorman, 2007).

73

13bit
it

/ / / 7: 52:31.. /c.
S

 

Table 6. Zurich Hazard Analysis Matrix of Severity versus Probability

SEVERITY
Critical

1 3
Probable 2 5
Occasional 4 6

Remote
Improbable 12 15

E
i
i

 

Note: this matrix includes a “Total Acceptance Risk Boundary Line”

ZHA emphasizes using peer consultation or expert panels to review the validity of
the assumptions and provide broad coverage of potential risks. The Zurich process is not
widely noted in peer-review publications, but one article in the Journal of Biotechnology
did discuss applications for it in biotechnology, and specifically, in biosafety. The article
discussed the usefulness of the matrix with the boundary conditions (Popping, 2002).

In addition, the United State’s Department of Labor, Occupational Safety &
Health Administration Office (OSHA) has approved Zurich Hazard Analysis “. . .as an
appropriate method for conducting the process hazard analysis which is required by
1910.1 l9(e)(2)” (Occupational Safety & Health Administration, 1995). The Animal and
Plant Health Inspection Service (APHIS) of the United States Department of Agriculture
(APHIS/ USDA) also recognized Zurich Hazard Analysis as a combination risk
assessment tool, meaning that the tool utilized a combination of systematic steps to

review and manage a risk (Charlton, 2004).
2.3.6.2 Example - MIL-STD-882D

In Mil-Std-8 82D, there is a general reference to formal risk assessments and
standards, but they usually only provided and outlined for the process and do not include

either the details of selecting characteristics or the quantitative values of those

74

Chara”

 

characteristics. The literature analysis and standards frequently emphasize the use of

experts and the need to take their insights into consideration in order to adapt the risk

assessment model for a specific situation. The Standard provides suggested mishap

probability levels that are flexible, based on the specific system being analyzed (see

Table 7).

Table 7. Suggested Mishap Probability Levels for Specific Events and Occurrences in Mil Std 882D

 

 

 

 

 

 

 

Description Level Specific Individual Item Fleet or Inventory

Frequent A Likely to occur often in the life of an item, Continuously experienced
with a probability of occurrence greater than
10"-1 in that life.

Probable B Will occur several times in the life of an item, Will occur frequently.
with a probability of occurrence less than 10"-

1 but greater than 10"-2 in that life.

Occasional C Likely to occur some time in the life of an Will occur several times.
item, with a probability of occurrence less
than 10"-2 but greater than 10"-3 in that life.

Remote D Unlikely but possible to occur in the life of an Unlikely, but can reasonably
item, with a probability of occurrence less be expected to occur.
than 10"-3 but greater than 10"-6 in that life.

Improbable E So unlikely, it can be assumed occurrence Unlikely to occur, but

 

 

may not be experienced, with a probability of
occurrence less than 10"-6 in that life.

 

possible.

 

Note: table adapted fi'om: (Mil-Std-882D, 2000)

With the Tylenol tampering example, a tampering did occur, and others have

occurred. The product does have, and will continue to have, a “frequent” probability

(Mitchell, 1989). The importance of calibrating the risk to the scale is clear. The use of a

very small data set, combined with the concept that any and every product will get

counterfeited somewhere in the world, could lead to defensibly picking any of the above

descriptions without something like the Counterfeit Product Risk Model. The probability

descriptions cover the life expectancy of the system, but the Standard offers flexibility in

terms of “unit of time, events, population, items or activity.”

75

 

 

The Standard also notes that “Assigning a quantitative mishap probability to a
potential design or procedural hazard is generally not possible early in the design
process.” This includes situations where there is little data or understanding of the
mishap risk. As long as the methodology is clearly defined, the Standard accommodates
deriving a probability fi'om “research, analysis, and evaluation of historical safety data
from similar systems.”

The recommended severity categories and values are in Table 8:

Table 8. Suggested Mishap Severity Categories in Mil Std 882D

 

Description Category Environmental, Safety, Health Result Criteria

 

Catastrophic I Could result in death, permanent total disability, loss
exceeding $1M, or irreversible severe environmental
damage that violates law or regulation.

 

Critical 11 Could result in permanent partial disability, injuries or
occupational illness that may result in hospitalization of at
least three personnel, loss exceeding $200K but less than
$1M, or reversible environmental damage causing a
violation of law or regulation.

 

Marginal III Could result in injury or occupational illness resulting in one
or more lost work days(s), loss exceeding $10K but less than
$200K, or rnitigatible environmental damage without
violation of law or regulation where restoration activities can
be accomplished.

 

 

Negligible IV Could result in injury or illness not resulting in a lost work
day, loss exceeding 82K but less than $10K, or minimal
environmental damage not violating law or regulation.

 

 

 

 

The importance of calibrating the Result Criteria to a specific risk is clear in cases
since most counterfeiting occurs in the “exceeding $1 million” range (N aim, 2005) and
thus, would always have a “Severity Description” of “Catastrophic.” Using the Tylenol
case as an example (seven deaths and an estimated $1 billion in economic impact), the
incident would have a severity rating of “I” (Catastrophic) whereas the Carver+Shock

shock scale was a 3-4 out of 10.

76

655

(am
11

u

3:.

 

A note is included with the original table that states “adaptation to a particular
program is generally required to provide a mutual understanding between the program
manager and the developer as to the meaning of the terms used in the category
definitions” (Mil-Std-882D, 2000). Also, it is mentioned that other techniques for risk
assessment may be more efficient. The Standard does not mandate exact following of the
Standard.

An example of the risk assessment matrix from the Standard is in Table 9.

Table 9. Example of mishap risk assessment values in Mil Std 882D

 

 

 

 

 

 

 

 

 

SEVERITY

Catastrophic Critical Marginal Negligible
>-‘
t: Frequent 1 3 7 13
d Probable 2 5 9 16
% Occasional 4 6 ll 18
8 Remote 8 10 14 19
sf Improbable 12 15 17 20

 

 

 

 

 

 

 

The mishap risk assessment values are grouped into mishap risk categories (Table
10). The definition or process used to define the categories is undefined in the standard.
The mishap risk acceptance level indicates the title of the government official who would
need to authorize accepting this risk in.

Table 10. Example mishap risk categories and mishap risk acceptance levels

 

 

 

 

 

 

 

 

 

 

Mishap Risk Assessment Mishap Risk Category Mishap Risk Acceptance Level
Value
1-5 High Component Acquisition Executive
6-9 Serious Program Executive Officer
10-17 Medium Program Manager
18-20 Low As directed
77

 

--
1...:

the V3

 

It is important to know that there is no discussion or guidance in selecting either
the values (1-5, 6-9, 10—17, 18-20) or the categories (high/serious/medium/low).

The Mil-Std-882D is especially applicable to the Counterfeit Product Risk Model
research because it is one of the most detailed and practical risk assessment guides. It is
also an accepted standard procedure for regulatory agencies such as FDA and USDA.
The FDA Combating Counterfeit Drug Task Force risk model development goals can be
achieved through this Standard. The Standard can be used to explicitly identify where

the Counterfeit Product Risk Model can add insight and value.

2.3.6.3 Example - Codex Alimentarius

The Codex Alirnentarious Commission (Codex) operates between the Food and
Agriculture Organization (F A0) of the United Nations (UN) and the World Health
Organization (WHO), and is charged with global food safety initiatives, including global
food standards. The Procedural Manual of the Commission specifically addresses risk
analysis; and its concepts are parallel to other agencies such as the U.S. FDA. The Codex
links components of risk assessment, risk management, and risk communication;
stressing science-based and transparent analysis; and an emphasis on the separation of
risk assessment and risk management to “ensure the scientific integrity of the risk
assessmen ” (WHO, 2004).

The Codex examines situations where “a risk to human health exists but scientific
data are insufficient or incomplete” and that “the degree of uncertainty and variability in
the available scientific information should be explicitly consider ” (WHO, 2004). In
addition, “risk estimates may be qualitative or quantitative, but should be quantitative to

the extent that is scientifically achievable” (WHO, 2004).

78

 

The Codex recommends, and offers discussions on, how to engage experts. As
with most of the other standards or regulations, the Codex does not provide product
specific or risk specific details that contribute to the detailed development of the

Counterfeit Product Risk Model.
2.3.6.4 Example - FMEA

Failure Modes and Effects Analysis (F MBA) is a quality control and risk analysis
system with underlying risk assessment and management concepts that are very sound
and insightful, although not suited to the all-encompassing counterfeit threat. FMEA is a
widely used, pro-active quality tool that focuses on design improvements and physical
failures (Kmenta & Ishii, 2000). The FMEA failure mode is defined here:

Failure Mode - the manner in which a component, sub-system, or system

could potentially fail to meet the design intent. The potential failure mode

could also be the cause of a potential failure mode in a higher level

subsystem, or system, or the effect of a lower level effect (Kmenta & Ishii,
2000).

FMEA focuses on system performance by analyzing reliability, maintainability
and safety (Onodera, 1997). FMEAS are most frequently used in early product
development and then again in manufacturing (Onodera, 1997). The FMEA system is
based on data gained from known recorded failures—in the lab and in the field— which
leads to the efficient use of a probabilistic approach (probabilities based on historical
data) (Kara-Zaitri, Keller et al., 1991). Other reports specifically identified FMEA in
manufacturing operations, focusing on “the ways equipment can fail or be improperly
operated,” with an emphasis on identifying the specific Single component that failed
(Graver, 2001). FMEA is used to analyze “risk by identifying hazards and suggesting

process design modifications” (Zambrano, Sublette et al., 2007). The FMEA emphasis

79

on monitoring recurring actions within a specific system is demonstrated in the abstract
of the FMEA Reference manual, published by the Society of Automotive Engineers
(SAE). Its focus is on “potential failure of a product/process” and identifying actions that
could reduce or eliminate the failure (Society of Automotive Engineers (SAE), 2002).

The key component of the F MBA is the Risk Priority Number (RPN), which is
used to assess the risk using the three criteria of occurrence, severity, and detection.
Detection focuses on identifying the failure before the customer receives the product.
Detection specifically looks at the physical product’s development, manufacturing, and
operations.

FMEA is a widely used quality and risk assessment process that is event and data
intense. It focuses on specific products and systems, so it does not directly apply to the

Counterfeit Product Risk Model.

2.3.6.5 . Example - ISO Standards

There are several International Organization for Standardization (ISO) standards
that are related to counterfeiting, but none apply directly or provide detailed insight for
the Counterfeit Product Risk Model (Branch, 2008).

ISO Standards for Risk

Of the one hundred and twenty active ISO Standards that include the term “risk”
in the title or abstract, there is only one active standard and only one proposed standard
that contribute to the Counterfeit Product Risk Model: “180/ IEC Guide 73, Risk
Management — Vocabulary,” which covers twenty-nine standardized risk vocabulary
terms (ISO, 2002, 2007b) and the pr0posed “ISO 31000 Enterprise Risk Management”

which covers general guidelines for enterprise risk management (ISO, 2007a, 2008a).

80

ISO/ IEC Guide 73 covers risk management terminology that has been incorporated in
the previous literature review (180, 2007b). ISO 31000 was under review at the time of
writing, with a publication date projected for 2009 (Standards New Zeeland, 2008).

The draft of “ISO 31010 Risk Management — Risk Assessment Guidelines,”
which has not yet been released publically, will include a more formal guideline on
assessment of risks than in ISO 31000 (ISO, 2008b). The Counterfeit Product Risk
Model would fit under section “3.29 Risk Indices,” because the nature of the data or the
emerging risk warrants using a qualitative or semi-qualitative comparative tool. The
standard discusses breaking the risk down to its risk components (such as probability and
severity) and to more detailed attributes. The standard discusses the use of qualitative
(“such as high, medium, and low”), semi-quantitative or quantitative methods based on
the “particular application, availability of reliable data and the decision making needs of
the organization.” The standard also discusses applying mathematical formulas to scale
the risk ranking.

As with the other standards reviewed, there is great detail on how to manage the
risk process, but little or no detail on when or how to evaluate whether an emerging risk
should even enter the risk analysis process.

These standards support the continuing definition and development of more
formal risk analysis and enterprise risk management but they do not specifically address
the counterfeit risk threat.

ISO 14971 Medical Devices — Application of Risk Management
ISO 14971, “Medical devices- Application of risk management to medical

devices,” is another excellent and effective risk management standard that supports many

81

of the concepts in the Counterfeit Product Risk Model, but, as with the other standards, it
does not offer the details necessary for building the risk assessment model.

ISO 14971 is important to healthcare providers because it supports risk
assessment for new equipment and procedures with an FMEA-type procedure (Christ,
2005; Dain, 2002). A key value of ISO 14971 is that it offers an “excellent structure for
analyzing risk” and balances “the risk of being too risk averse” and the “the risk of being
too risky” (Fotis & Bix, 2006).

As with many of the other ISO type standards and processes, ISO 14971 was
developed to reduce hazards and harm in specific system failures and for specific
products in specific applications (Rudolph, 2003). To that end, roughly 80% of
manufacturers use F MBA, to some extent, in order to implement ISO 14971 (Schmidt,
2004). F MBA is very effective at evaluating and controlling the manufacturing and
operations processes, and a fault tree analysis is excellent for considering risks in
handling (Schmidt, 2004).

For the products covered by ISO 14971, this quality risk management approach is
effective, in part, due to “huge amounts of data — data availability;” “data in separate,
diverse systems;” “huge amounts of documentation;” “documentation to support
submissions;” and “production records.” (Johanning & Dahlgaard, 2007). These
attributes are in limited quantities or not available in product counterfeiting.

“[In ISO 14971] risk analysis identifies potential foreseeable hazards derived
fi'om product testing, complaints, adverse events, recalls, audit observations, and other
product or process deviations that are involved with the use, misuse, or abuse of the

product” (Smith, 2004). Much of the literature focuses on potential harm to humans or

82

hazards from products or product defects, or from user errors, and reasonably foreseeable
misuse and abnormal use (Murphy, 2006).

As with other standards, such as HACCP and GMP, it would be logical for those
implementing ISO 14971 to consider the counterfeit risk in its hazard identification step.
Once a specific type of counterfeiting is identified that has a specific risk, this specific
risk could be evaluated using a fault tree analysis (or other method), and then by working
backwards, processes or systems that would minimize the hazard or harm could be
implemented (Clinton, 2008).

ISO Standards for Security and Traceability

The ISO standards that include “security” or “traceability” in their title or
abstract, or that are noted in other literature as related to product counterfeiting, are
(Branch, 2008):

ISO 27000 Information Security Management Systems

ISO 28000 Supply Chain Security Management

ISO 22000 Food Safety and Security

ISO 14000 Environmental Management & Security

ISO 17799 Generic Information Security Standards for Computers and
Information Technology

These offer insights or perspectives on how risk models are used in other areas,

but they do not contribute to the Counterfeit Product Risk Model.
2.3.6.6 ICH Q9

The International Conference on Harmonization (ICH) has created a quality
standard, identified as Q9 (International Conference on Harmonization, 2005), and wide
adoption of the quality standards that Specifically address risk management have occurred
(Wechsler, 2007). The ICH has developed a range of programs—such as Q8,

9

“Pharmaceutical Development” and/or Q9 “Quality Risk Management’ —which have

83

been adopted by the EU, Japan, and the US (FDA) (CBER/FDA, 2006). The goal of ICH
is to harmonize work processes and build off current regulations and industry practices
such as GMPS and standards such as ISO 9000 (a quality management standard) (Martin,
2007). In combination, Q8 and Q9 focus on the overall pharmaceutical manufacturing
process. A Q10 program is also underdevelopment which will directly link standards Q8
and Q9 (W echsler, 2007).

The Q9 standard addresses risk management and specifically focuses on “risk-
management principles applicable to drug development and manufacturing. It includes
methods such as hazard analysis and critical control points, and failure-mode effects
analysis [FMEA]” (Bush, 2005). The standard defines terms and includes conceptual risk
models. The standard refers to quality risk management systems (QRM)—which is the
traditional risk analysis of probability versus severity—to quantify an overall risk
(W echsler, 2007).

The ICH Q9 (and Q8 and Q10) build off of existing quality and risk management
processes and best practices (International Conference on Harmonization, 2005). While
the ICH standards do not mention counterfeit or economic fi'aud, one aspect of ICH Q9
that has an application to this research, is that it includes “detect-ability” in the formula
with probability and severity. The key insight is that counterfeit products are detectable
if there is a process to look for them and incorporates some component or method to
actually identify the fakes.

With the ICH Q9 focus on manufacturing and operation processes and risks, it
includes many of the same excellent insights and limitations, of the other standards.

However, they do not directly apply to the Counterfeit Product Risk Model.

84

2.3.6.7 Other Risk or Quality Programs

Several other risk and quality programs or standards that have been reviewed in
the analysis and design of the Counterfeit Risk Model are discussed in the following:

The Preliminary Hazard Analysis (PHA) process considers “prior experience or
knowledge of a hazard or failure to identify future hazards” and—although it is not
specifically intended as a tool or model for counterfeiting—product counterfeiting could
be included as a hazard (International Conference on Harmonization, 2005).

The Hazard Operability Analysis (HAZOP) program focuses on design and
operating deviations that lead to risk. There is no mention of counterfeiting or fraud. If
the process were to be extended to suppliers or contract manufacturers, the counterfeit
product risk could be incorporated into this process (International Conference on
Harmonization, 2005).

The Failure Mode, Efl”ects and Criticality Analysis (FMECA) is different than
F MEA, since it includes more consideration of the severity of an event. As with FMEA,
FMECA focuses on reducing risk through product design or manufacturing
specifications. Product counterfeiting could be included as a potential source of failure or
variation, but it is not a tool or model that directly applies to the Counterfeit Product Risk

Model (International Conference on Harmonization, 2005).

2.3.7 Industry and Agency Reviews

This information is included because many of the basic concepts could apply to

the Counterfeit Product Risk Model.

85

2.3.7.1 Hurricane-Related Risk Assessment

Hurricane-related risk assessment is covered in the Society for Risk Analysis
Journal of Risk Analysis (Jain & Davidson, 2007). In this research a model of the wind
risk to wood-flame houses was developed, in order to analyze how different building
variables would affect direct economic loss. Variables included changes in building
codes and population growth that were correlated to the relative strength of the
hurricanes. To gauge the severity of the wind, historical storm data was used in a
classical statistical Monte Carlo simulation to judge the risks. This analysis is possible
(there is data on the winds in a hurricane) and logical (the winds act consistently) since it
is based on specific atmospheric conditions.

For the Counterfeit Product Risk Model, this article emphasizes the value of
model development, even though there may be a very small sample size. It supports

using qualitative estimates of probability.
2.3.7.2 Risk Assessment in Criminal Justice

This overview of risk assessment in criminal justice was developed to investigate
those aspects and tools that are applicable to the Counterfeit Product Risk Model. The
academic discipline of Criminal Justice has incorporated risk prediction since the 1930’s,
and has evolved fi'om the use of PAT (process analysis triangle of criminal] victim/
opportunity) to the more modern and more practical, Situational Crime Prevention work
(Clarke, 1997). Other processes that have been developed include RAT (routine activity
theory), rational choice, and SARA (Scanning, Analysis, Response and Assessment)

(Center for Problem-Oriented Policing, 2006; Felson, 1998; Town, 2001 ).

86

Goldstein developed eleven basic elements behind problem-oriented policing that
were referenced in Bullock, and that have applications in an anti-counterfeit strategy.
These are basic steps in a Counterfeit Product Risk Model (Bullock & Tilley, 2003;

Goldstein, 1979).

Eleven ‘basic elements’ of ‘problem-oriented policing’

grouping incidents as problems

focusing on substantive problems as the heart of policing

effectiveness as the ultimate goal

the need for systematic inquiry

disaggregating and accurately labeling problems

analysis of the multiple interests in problems

capturing and critiquing the current response

an uninhibited search for a tailor-made response

adopting a proactive stance

0. strengthening the decision-making processes and increasing
accountability

11. evaluating results of newly implemented responses

”99°99‘99pr

In his book “Understanding Risk in Criminal Justice,” lKemshall discusses two
criminal justice concepts which are applicable to the Counterfeit Product Risk Model
(Kemshall, 2003). The first is that both first generation (subjective assessment,
professional judgment, etc.) and second generation (systematic, empirically grounded,
etc.) risk assessment tools should be used when reviewing the counterfeit threat
(Kemshall, 2003).

The second important concept examines how rational choice theory may apply to
criminals and counterfeiters. For instance, by raising costs and lowering benefits, the
economics of the opportunity can rationally deter criminals (Kemshall, 2003).

A key application to the Counterfeit Product Risk Model is that while “Criminal
Justice” uses extensive risk prediction, such as PAT and Situational Crime Prevention,

there are no standardized, specifically focuses or precise tools.

87

Key findings of this research which are applicable to the overall anti-counterfeit

strategy:

Criminal “opportunities” will greatly increase because the value of the opportunities
are increasing along with the criminals’ advancing manufacturing capabilities and
increased technical skill, which they will continue to find ways to optimize.

By definition, “pure risk” has no upside whereas “speculative risk” does, but the pure
risk to the attacked retailer provides a possible positive speculative risk for the non-
attacked retailers (Nader, 2002). For example, if the attacked product is a necessary
consumable, such as infant formula, consumers will move to some other trusted
retailer to satisfy this demand. As long as there is still an opportunity, the criminals
are very adaptable (Kemshall, 2003).

While it is important for retailers and manufacturers to support increased legislation
and international enforcement, the immediate benefits are from actions to implement
the keys of Situational Crime Prevention of ‘target hardening’ and ‘surveillance’

within their own companies (F elson & Clarke, 1997).

2.3.7.3 Review of Risk Assessment in US Forest Service Wildfire

Danger Rating Level System (WFAS)

This review of risk assessment in the US Forest Service Wildfire Danger Rating

Level was done to determine if any of the aspects and tools are applicable to the
Counterfeit Product Risk Model. In fact, this risk assessment system may actually be the
most directly applicable to the Counterfeit Product Risk Model. The Wildfire Danger
Rating Level is a successful, clear, and consumer-recognized danger rating system (US

Forest Service, 2006).

88

The rating system is comprised of scientific measures, taken by assigned experts,
that take into account current and “antecedent” weather (1 -10-100-1000 hour future
weather or impact of the weather on the forest) (US Forest Service, 2006). The rating
system normalizes rating classes across all environments and conditions. A Fire Potential
Index (F PI) (Equation 1) is derived from the scientific measures in a formula, and is then
applied, using an “adjective class rating” method. The rating appears to be randomly set
at quadrants, but the input variables are weighted. For example, “The value 35(/ 100) is
used as the lowest maximum percent green [there are higher color categories], even for
arid areas of the West” (Burgan, 2006). The FMI formula is statistically reviewed and if
the r"2 value for all years was 0.72, then the “correlation between the FPI and fire

occurrence was very high” (Table 11 (Burgan, 2006).

Table 11. Fire Danger Rating and Color Codes from the WFAS System

 

 

 

 

 

 

 

 

 

 

 

Fire Danger Rating Color Code Approximate FPI Index (applied to
percentages) (0 = no data)

Low (L) Dark Green 1-20
Moderate (M) Light Green or Blue 21-50
High (H) Yellow 51-70
Very High (VH) Orange 71-90
Extreme (E) Red 91—100

FPI=(1-TNf)*(1- LR)*100 (1)

Where equation [1] produces FPI values that can range from 0 to 100. The
FPI will equal 0 when the TNf is 1 (the dead fuel moisture equals the
moisture of extinction) or the LR value is 1 (the vegetation is firlly green).
These circumstances do occur, but the FPI is limited to a minimum value
of 1 so that areas outside the United States can be identified as the value 0
(no data). The FPI will attain a value of 100 if the LR is 0 (all the
vegetation is cured) and the 10 hour time lag fuel moisture is at its
minimum value of 2 percent.

89

 

Beyond the simple warning for citizens, the Fire Danger Rating is used to allocate
funds, personnel and firefighting equipment (Burgan, 2006). The objective of the process
is to attain 97-99% confidence that a “large” fire will be suppressed in 1 to 3 days.

The key findings of this research which are applicable to the Counterfeit Product
Risk Model are:

0 US Agencies and citizens are familiar with risk danger rating systems

0 The Danger Ratings are tied to some specific actions or procedures that are
implemented when a rating changes

0 Reduction in variability and uncertainty in measurements (which would initiate
changes in preparedness). The variables are scientifically derived and measured by
trained experts.

- A simple system of measurements and derivation can be extremely effective in
predicting and managing specific risks—in fact, simple systems of derivation and
output seem to be very effective.

0 As the danger rating experience and expertise grow, the system includes early
warning signs that might lead to future risky conditions. In the case of fires, the
change in vegetation over time (how much will the vegetation dry or moisten over 1-
l 0-100-1 000 hours) is considered.

a Once there is a large enough data set for statistically significant results, behaviors can

be changed based on the application of classical statistical tools.

90

2.3.7.4 Airplane Birdstrikes

The process for predicting the risk of birdstrikes to airplanes, exhibits many
similarities to counterfeit risk prediction. This section is based on an article in the
Journal of Risk Analysis published by the Society for Risk Analysts (Allan, 2006).

Airplane birdstrike events are “not uncommon” (750 strikes in the UK per year)
and the results are perceived to be costly, though quantitative cost estimates are not
available in the public domain. This analysis converts quantitative measures (number of
hits) into a qualitative probability judgment of low/med/high, then applies the probability
data to the probability x severity matrix to arrive at a final, overall qualitative risk
assessment (Allan, 2006).

A particular focus of this analysis is to assist efficient risk mitigation, by helping
to evaluate the actions that will provide the greatest benefit. This process is being used in
the United States, Canada, Australia, and New Zealand.

The application to the Counterfeit Product Risk Model derives fi'om its use of the
probability x severity matrix, and use of qualitative measures of probability for small data
set analysis.
2.3.7.5 US Government Accountability Office (GAO) Reports

The US Government Accountability Office (GAO, before 2004 known as the US
General Accounting Office) conducts investigations and research to answer questions

posed by other government agencies. One such report is “RISK MANAGEMENT -
Strengthening the Use of Risk Management Principles in Homeland Security” (United

States Government Accountability Office, 2008b).

91

 

Topics related to the Counterfeit Product Risk Model have been reviewed, such as

risk management, intellectual property rights, economic fraud, food safety, terrorism and

even counterfeit product in use in nuclear facilities. The most applicable reports are

listed here (listed by most recent data):

1)

2)

3)

4)

5)

6)

7)

8)

9)

RISK MANAGEMENT - Strengthening the Use of Risk Management Principles
in Homeland Security, United States Government Accountability Office (GAO),
GAO-08-904T, June 25, 2008

INTELLECTUAL PROPERTY, Federal Enforcement Has Generally Increased,
but Assessing Performance Could Strengthen Law Enforcement Efforts, United
States Government Accountability Office (GAO), GAO-08-157 March 2008,

INTELLECTUAL PROPERTY - Risk and Enforcement Challenges, Statement of
Loren Yager, Director International Affairs and Trade, United States Government
Accountability Office (GAO), GAO-08-177T, October 18, 2007

FEDERAL OVERSIGHT OF FOOD SAFETY - High-Risk Designation Can
Bring Attention to Limitations in the Govemment’s Food Recall Programs,
Statement of Lisa Shames, Acting Director Natural Resources and Environment,
Testimony, United States Government Accountability Office (GAO),Tuesday,
April 24, 2007

INTELLECTUAL PROPERTY - Better Data Analysis and Integration Could
Help U.S. Customs and Border Protection Improve Border Enforcement Efforts,
United States Government Accountability Office (GAO), GAO-07-735, April
2007

INTELLECTUAL PROPERTY - Initial Observations on the STOP Initiative and
U.S. Border Efforts to Reduce Piracy, Statement of Loren Yager, Director
International Affairs and Trade, United States Government Accountability Office
(GAO), GAO-06-1004T, July 26, 2006

INTELLECTUAL PROPERTY - U.S. Efforts have Contributed to Strengthened
Laws Overseas, but Significant Enforcement Challenges Remain, Statement of
Loren Yager, Director, International Affairs and Trade, United States Government
Accountability Office (GAO), GAO-05-788T, June 14, 2005

INTERNET PHARMACIES - Some Pose Safety Risks for Consumers, US
Government Accountability Office (GAO), GAO-04-820, June 2004

CIGARETTE SMUGGLING - Federal Law Enforcement Efforts and Seizures
Increasing, United States General Accounting Office (GAO), GAO-04-641, May
2004

92

10) FRUIT JUICE ADULTERATION: Detection Is Difficult, and Enhanced Efforts
Would Be Costly, United States General Accounting Office, GAO GAO/RCED-

96-18 November 1995

11) NUCLEAR SAFETY AND HEALTH, Counterfeit and Substandard Products Are
a Government-wide Concern, Nuclear Safety and Health (GAO/RCED-91-6
Counterfeit and Substandard Products), United States General Accounting Office

(GAO), October 1990

Although the reports offer insight and direction in assessing the counterfeit threat,
the reports do not define models or tools, so there is no direct contribution to the
Counterfeit Product Risk Model.
2.3.7.6 Value At Risk (V aR)

The banking and capital management industry concept of “Value at Risk” applies
to the Counterfeit Product Risk Model. This section is based on the book “Risk and Bank
Capital Management” by Francesco Saita (2007) (Saita, 2007).

Value at Risk (V aR) is the amount of a corporation’s value, in terms of actual
dollars, that is at risk during normal business operations, such as the loss on a sale of an
investment or a bad loan. A critical limitation of VaR—both for this industry and for
anti-counterfeit strategy—is that it does not include other losses, such as brand equity,
civil liability, etc. Nevertheless, VaR does include risk management concepts that clearly
apply to the Counterfeit Product Risk Model.

To compare risks corporation-wide, the risks need to be able to be compared in
the same way; in a harmonized risk system. Most risk management actions require
financial expenditures, and analysis should be presented in those terms. When applied to
new or emerging risks, there is a known difficulty in deciding what risk to consider in the
first place, how to calculate these risks, and how to aggregate them to a corporate-wide

level. “In summary, while risk measurement may be considered a science, risk

93

 

management, capital management, and capital allocation remain largely a blend of
science, experience, and art” (Saita, 2007).

VaR focuses on operational risk and the inclusion of “external events” in the
analysis applies to the counterfeit risk. “Operational risk is defined as the risk of loss
resulting from inadequate or failed internal processes, people and systems or from
external events. This definition includes legal risk, but excludes strategic and
reputational risk” (Saita, 2007).

In measuring the risk, as with other systems, there is an emphasis on achieving
quantitative measures. Scenario analysis is recommended as a way to begin to
understand, map, and start to quantify these new risks, using the probability versus
severity matrix. The early steps of model development or risk management should be
made based on historical experience, if available and reliable, and then correlated with
expert opinion.

Scenario analysis is used to develop forward-looking risk models in an effort to
learn about the underlying processes that lead to the risk. The forward-looking processes
use key risk indicators (KRIS) to identify and anticipate potential risks. There are several
key attributes of KRIS that apply to the Counterfeit Product Risk Model (Saita, 2007):

1) strongly related to the frequency and/or severity of critical operational

2) highrsedundant [the risk should only be counted once],

3) as measureable and easy to monitor as possible, and
4) auditable, i.e. subject to easy verification.

Many authors and articles support the view that VaR is limited for considering
extreme events (‘tail risk’ or ‘fat tail’ events—extreme events that are very infrequent but
have potentially very high impact) in the calculation (Quaranta & Zaffaroni, 2008;

Tibiletti, 2008; Wang, Chaudhury et al., 2008). Hallerbach (2004) discusses

94

“decomposing downside risk into components that are attributable to each of the

underlying risk factors” (Hallerbach & Menkveld, 2004). In Duffie & Pan (1997), a

“lognorrnality assumption” is presented, which emphasizes unaccounted-for risk when

assuming the reality will match the expected prediction of activity.

The bank and capital management risk management concept of Value at Risk has

many insights valuable to the Counterfeit Product Risk Model:

There is an acceptance that hazard identification has already occurred; the risk is
already identified and considered important enough to run through the risk
management process. This is a recognized gap in many of the risk management
models and processes.

An interesting insight is that some risks seem to be so rare that risk mitigation (trying
to reduce the frequency or severity) steps are difficult or impossible to quantify, so
risk management moves fiom mitigation to acceptance.

Scenario development is valuable in forward-looking models for new or emerging
risks.

The risks should be quantified in terms that can be harmonized in corporate-wide risk
management systems; in this case, to facilitate the effective allocation of capital.

The beginning steps of risk management are as much science as art.

The first step in quantifying a new or emerging risk is to map the underlying
operational processes that lead to the risk, and to begin to develop the frequency
(probability) and severity estimates.

Risk management should be based on historical experience, if reliable and available,

and correlated with expert collaboration.

95

2.3.8 Risk Model Research Summary

The options for conducting risk assessment are vast. A close review of actual risk
assessment models, including the standards that are in place, demonstrates that the
operational risk management (ORM) or “probability versus severity” matrix is commonly
recommended and used.

Most of the models and processes use some quantitative measures but end up
using qualitative statements of probability, severity, or of overall risk. The range of
models or processes that were defined usually breakdown an overall risk into a series of
mutually exclusive, comprehensively exhaustive (MECE) factors. Some models or
processes break the main factors down to more detailed sub-factors. In other models,
there is already a more detailed analysis that could be considered sub-factors.

Although there are many aspects of models or processes that could contribute to
some aspects of the Counterfeit Product Risk Model, the literature analysis and report

review did not identify any directly applicable models.

96

3.0 MODEL DEVELOPMENT

This section covers the methods used to develop the Counterfeit Product Risk
Model, the series of propositions that was presented to an expert panel of industry
experts, and an industry survey that was presented to industry professionals to review
current industry actions and trends. The propositions and the model concepts were tested
in the Expert Panel Survey (Chapter ##) and a practical example of the use of the model

is included in the Case Study (Chapter 5).

3.1 Model Overview

Before presenting the Specific Counterfeit Product Risk Model details, an
overview is provided here. The model focuses on the probability portion of the general
probability versus severity matrix. The probability is based on a series of five factors
which, in turn, are based on a series of three (or more) sub-factors. The five factors are
proposed to be used with all counterfeit products, whereas the sub-factors are to be
developed for the specific products being reviewed.

Figure 3 is an overview of the Counterfeit Product Risk Model. The overview
shows how the sub-factor derivation leads to the ranking of the sub-factors, which feeds
up to the ranking of the factors, which finally feeds into the overall ranking of
probability. The Counterfeit Product Risk Model uses a cascading approach to derive the
sub-factors, which cascade up to the factors, which cascade up to the overall risk rank.
Vice versa, by cascading down from the overall risk rank to ever increasing orders of
detail can facilitate understanding between the parties developing or using the

information. This concept is similar to the model cascading approach discussed by

97

Hovvick (2008) in “Building confidence in models for multiple audiences: The modeling

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cascade.”.
Counterfeit Product Risk Model:
Overall Risk Rank

1 r ! I fiL

Counterfeit Counterfeit Counterfeit Counterfeit Market
History Ability Attractiveness Hurdles Profile

Sub-Factors Sub-Factors Sub-Factors Sub-Factors Sub-F actors

I 1 1 1 1

2 2 2 2 2

3 3 3 3 3

l l !
Sub-Factor 3 Derivation Sub-Factor 2 Derivation Sub-Factor 1 Derivation
Define how the rating Define how the rating Define how the rating
was derived was derived was derived

 

 

 

 

 

 

 

 

Figure 3. Counterfeit Product Risk Model Overview

Howick emphasizes the cascading model building process, specifically called the
four stage reversible cascade, as being transparent, auditable, and clear. The cascading
Process has many advantages such as: achieving comprehensiveness, developing
organizational learning, testing the veracity of multiple perspectives, modeling
transparency, achieving common understanding across many audiences and promoting
conficlence in the [organization’s understanding of all the risks included in] models”
(HOWick, Eden et al., 2008).

I-lowick’s model uses a series of steps to convert “natural language to numerical
Sin“nation and back again.” Essentially taking real world risk attributes, and through

8
e"’Sfi‘al steps, compounding the concepts to a single model.

98

3.2 Research Introduction

The Counterfeit Product Risk Model was developed using literature sources,
agency report reviews, and extensive peer consultations with industry, agencies,
associations, and other academics. The literature supports this early engagement of
experts, formally called “peer input” or “peer consultation” (Patterson, Meek et al.,
2007). The Model has been presented and/or discussed with the members of a wide
range of groups including: Michigan Department of Agriculture (MDA), National Center
for Food Protection and Defense (N CFPD), International Association of Packaging
Research Institutes (IAPRI), Society for Risk Analysis (SRA), International Association
for Food Protection (IAF P), International Authentication Association (1AA), American
Society of Industrial Security (ASIS), and others.

Prediction of counterfeiting risk is such that classical statistical models or
formulas are not effective. For theoretical purposes, formulas are often developed to gain
general insight on how entire systems function, but that work is not useful in developing
a specific product Counterfeit Product Risk Model. If classical statistical models or
formulas were used, the heuristic (derived from personal experience rather than actual
data) nature of the root data would lead to either the impression of excessive precision or
confidence intervals so wide that the outcome would be unusable for specific risk
management decisions.

During the literature analysis and in discussion with the industry and agencies, it
became clear that there was very limited public information related to counterfeit risk, or
even the very first steps of hazard identification, a critical part of a counterfeit risk

assessment. This insight led to a modification of the original research concept, in a way

99

that led to the development of a more fundamental propositions that would assist future
researchers in more detailed, product, or industry specific risk assessments (C-STAT,
2008c). Propositions are word descriptions of the underlying premises of the concepts
rather than formulas (F eys & Fitch, 1969; Hazewinkel, 1988). Thus, basic propositions
were developed through literature and report review, and peer consultation.

A proposition is a preliminary step in the creation of a more formal research
project or hypothesis (Cramer & Howitt, 2004; Eisenreich & Sube, 1982; Weisstein,
2003). Propositions are usually very basic in nature and reference scholarly sources.

The series of propositions that were developed were used to create the
experimental design for this research. Propositions are often included as preliminary
work before formal hypothesis are developed (Cramer & Howitt, 2004; Eisenreich &
Sube, 1982; F eys & Fitch, 1969; Weisstein, 2003). Very often, propositions are used in
series, with each proposition building on the previous propositions (Karush & Tarcov,
1989)

The propositions were supported by a series of comments referred to in this
research as a “Strategy Statement.” These propositions, and their supporting “Strategy
Statements,” were developed to methodically review seminal concepts in the

development of the Counterfeit Product Risk Model.

3.3 Propositions

3.3.1 Proposition 1

Proposition 1 — The probability versus severity matrix is a logical starting point for the

Counterfeit Product Risk Model.

100

Concept Statement 1-1: A model was developed, and supported, using the
probability versus severity matrix that is in use by the FDA and others (Figure 4). These

points support the Concept Statement.

  

Probability

     
     
  

   

Medium

    

Very High Low Very Low

 

High

 

Very High
High

Medium

 
   
   
   
 

Low

Very Low

  

Figure 4. Example of Probability versus Severity Matrix

0 Several models were found in the literature but were not supported by scholarly
research. The most applicable model was the FDA’s Operational Risk
Management (ORM) model and the associated risk matrix. This is identical, in
principle, to the Zurich Hazard Analysis concept (Moorman, 2007).

o The operational risk tools (e.g. ANSI, ISO, EU, Mil-Std, etc.) are excellent
process for setting up the model but do not have scholarly support and do not
provide any guidance relative to counterfeit risk.

0 Classical statistical tools cannot be applied to the counterfeit concept at the hazard
identification stage, since the data sets are small and/or incomplete (though
statistical tools can be applied to incomplete or non-inclusive data sets for a

theoretical exercise) (Claycamp, 2006; Jablonowski, 1996).

101

The literature frequently reinforces the concept that the model developer must
understand what decisions will be made with the model, the resources available
and who will be using the model.

Concept Statement 1-2: The use of qualitative ranking in the model was

supported by the following points.

The operational risk management models use qualitative characteristics
(introduced in Section 2.2) such as high/medium/low (Buchanan, 2007). The
model development procedures recommend a simple format for early hazard
identification, using the available data and for ease of use (increasing the chance
they will be used).

There has been support over many years of using “value” words rather than
numbers for communicating risk (F eys & Fitch, 1969; Hazewinkel, 1988;
Zambrano, Sublette et al., 2007).

If the factors or sub-factors were ranked or weighted quantitatively, the heuristic
nature of the decisions, or the incomplete nature of the data could lead to what the
USDA has called “excessive precision” (Hassenzahl, 2006; Pittinger, Brennan et
al., 2003).

The literature included methods using three to nine value word categories. For

counterfeit product research, the extreme tails of the data were defined as very low and

very high, to represent close to the 0% and 100% extremes. For counterfeiting and

diversion, there is always some possibility that an event will or will not occur. Using

high/medium/low is very common, and with the counterfeit risk, it is important to

emphasize that there is always a risk for the extremes (very small but not 0% and very

102

large but not 100%), so very high and very low were included. To cover two extreme
tails and a common center of high/medium/low, it was most appropriate for the
Counterfeit Product Risk Model to evenly cover the range with five risk rankings from
very high (VH), high (H), medium (M), low (L), and very low (V L). The specific risk
ranking (VH, H, M, L, VL) values for the qualitative measures (dollars of loss, number of
injuries, etc.) are to be defined by the users of the model.

Further breakdown of the risk rankings was not efficient given the uncertainty of
the data and the heuristic nature of the risk categorization. Also, there is no consistent
method to determine the break points between the qualitative categories.

Concept Statement 1-3: No other models or processes were found that better
supported the research goal, including Value-At-Risk (financial) to the Forest Service
Wildfire Danger Rating System, including the FDA Operational Risk Management and
Carver+Shock models. The following supports the Concept Statement.

0 While Value-At-Risk has many concepts that can be applied to the Counterfeit
Product Risk Model, a critical limitation is the explicit focus on clearly
quantifiable financial loss fi’om a risk position. The concept does not consider
tangent exposure such as brand equity or civil litigation.

0 The FDA Carver+Shock model is commonly used in food defense risk assessment
research (CFSAN/F DA, 2007a). Carver+Shock does not directly apply to the
Counterfeit Product Risk Model since it was built on the premise that there is a
malicious intent, or more specifically, intent to cause shock or terror. The
counterfeiter, on the other hand, is seeking economic gain rather than causing a

physical attack—the public health threat is through negligence.

103

0 Many other models are used in risk assessment. A list of those specifically
reviewed are as follows:

0 FDA: ORM, Carver, Carver+Shock

0 Food Industry: HACCP, GAP, GMP, ALERT, FSIS/CF SAN, Codex
Alimentarius/ UN-FAO, etc.

0 Operational Risk: Mil-Std-882D, ANSI B11 TR3, 18014121, EN1050,
etc.

0 Others: Zurich/ Kellogg’s, airplane bird strike, hurricane building codes,
CJ/ situational crime prevention, Forest Service Wildfire Danger Rating
level system, etc.

Concept Statement 1-4: The focus on only the probability portion of the risk
matrix was supported by the following points (see figure 4). These points support the
Concept Statement.

0 Peer consultation supports that a detailed review of the probability part of the
probability versus severity matrix would help the risk assessment process.

0 Severity is calibrated using an organization’s own, unique, risks. For example,
the severity of failure events would be quite different for a writing pen product

and an airplane part company.
3.3.2 Proposition 2A

Proposition 2A — It is logical to break down the overall Counterfeit Product Risk Model

to a mutually-exclusive, comprehensively exhaustive set of factors.

104

Concept Statement 2A-1: The use of mutually-exclusive and comprehensively-
exhaustive factors was supported by the following points. This point supports the
Concept Statement.

0 The previously covered peer consultation and the literature review supported

dividing the factors into groups that were mutually-exclusive (characteristics

clearly fall into one factor) and comprehensively exhaustive (all major factors

are included).

3.3.3 Proposition 2B

Proposition 28 — It is logical to use these factors in the Counterfeit Product Risk Model:
Counterfeit-History, Counterfeit-A bility, Counterfeit-A ttractiveness, Counterfeit-Hurdles,
and Marketing Profile.

Past evidence of counterfeiting or diversion is the most accurate predictor of
future counterfeiting or diversion (deKieffer, 2006b; FDA, 2004; INTERPOL, 2007;
WHO, 2007b). Data confirming that the product has been counterfeited is evidence that
the product has been counterfeited, but no record of counterfeiting is not evidence that
the product is not counterfeited (deKieffer, 2006b; Liang, 2006).

Thus, in following the logic outlined in the FDA Combating Counterfeit Drugs
report (as well as other such as the WHO and COE) (Council of Europe, 2006; FDA,
2004), this model expands from the known infiingements to additional root factors.
Other root factors were reviewed in the literature. The five root factors (above) were

chosen to cover the full range of counterfeit issues.

105

Concept Statement ZB-l: Factors beyond the known infiingements (counterfeit
and diversion) will be included to provide more detail in the model. These points support
the Concept Statement.

0 This model considers the known infiingements and beyond to examine

additional root risk factors. This step is important in ranking risks within a
group of known counterfeited products or to rank a new product versus a
known counterfeited product. The top predictor of a counterfeit or diversion
risk is past evidence of the risk. Knowledge of past infringement is included
as a predictor in the FDA Combating Counterfeit Drugs report (FDA, 2004)
and in the FDA (food) Carver+Shock risk modeling software questionnaire
(CFSAN/FDA, 2007a).

0 To enable additional flexibility, the model can be used without the known
counterfeiting or diversion factor for a product or a product category. This is
effective if a sample of products may include some that have known
infiingements and others that do not have evidence of known infiingements —
if known infiingements were included as a factor, the known infiingement
products would always score a higher Counterfeit Product Risk Model.

Concept Statement 2B—2: The full range of factors includes: Counterfeit-History,
Counterfeit-Ability, Counterfeit-Attractiveness, Counterfeit-Hurdles, and Market Profile.
These points support the Concept Statement.

0 This research has resulted in two sets of root factors that are intended to cover the

full range of factors (mutually exclusive and collectively exhaustive) that lead to a

counterfeit risk.

106

0 Known infiingements - past history of counterfeiting or diversion of the
product or the product category (same type of product but another brand)

0 Root factors — specific for the product and supply chain

- The root factors are:

0 Counterfeit-History (history of counterfeit or diversion)

O Counterfeit-Ability (easy to copy, easy to dupe the consumer or customs)

O Counterfeit-Attractiveness (value to volume ratio, profitability per unit)

0 Counterfeit-Hurdles (security, countermeasures)

0 Market Profile (for example, the consumer demographics and
psychographics, legitimate supply chain, illicit supply chain, use of
contract manufacturers, sales process, etc.)

- The characteristics of the product and supply chain (authorized repackaging,
internet sales, secondary wholesalers, consumers seeking non-deceptive

counterfeits, etc.) are included in the Market Profile root factor.

3 -3 -4 Proposition 2C

Proposition 2c — It is logical to use a mathematical summary fimction to total the sub-
factors into one overall factor value, or several factors to one overall probability value.
Each matrix includes a “Summary Function” (for example, A+B+C+D=D) which
enables the multiple factors to be totaled into one overall value. The weighting of the
factors should be determined by the specific user since each set of factors is potentially
uni 9‘13 for each product, company, and industry.

In the “Function to Summarize to Total” in Figure 5, the variable a represents

th
e probability rank (V H, H, M, L, VL) of the first factor (1. Counterfeit-History), and so

107

 

on for variables “B” to “E.” Variable “F” is the probability rank for the “Total Rank,”

which was the average of the probability rank of the five factors.

The “Function to Summarize to Total” formula, weighting, and “Total Rank” are
detertnined by an Expert Panel. Considering the previous review of the nature of the data
and the warnings about keeping the analysis useable for the modelers, using this manual
(heuristic) process is supported, and preferable for few experiments.

The formula can use common spreadsheet “Logic Function” combined with a
“Matching Text” function to automate the “Total Rank.” For example, the logic function
earl identify the rank (since the “X” would be in one of the VH, H, M, L, VL categories)
and feed that into the formula as a binary number (0, 1). An example of using the

automated, mathematical model based “Total Rank” derivation is included in Appendix
E -

The process for developing the Summary Function should be generated by an

exPert panel, and the process documented. Each factor was defined by factor specific

Sub— factors which are presented in Figure 5.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Probability
Very . . Very
Hi ! High Medrtun Low Low
TOtal Rank x
W Function (for example) L A+B+C+D+E=F ]
l -
/C\°untcrfcit-History X
2 -
3 Counterfeit-Ability x
4 ‘ Counterfeit-Amctiveness x
I S ‘ :Ollnterfeit-Hurdles X
\ Ial‘ket Profile X
Flgu
re 5. Example of Factors Defining Probability for the Overall Counterfeit Risk

108

 

Concept Statement 2C-1: It was logical to use a mathematical summary function
to total the risk to one rank These points support the Concept Statement.

- Each matrix includes a “Function to Summarize to Total” (for example,
A+B+C+D=E) which enables the multiple factors to be summarized into one
rank. It was recommended that this function be determined by the specific user
since each set of factors was potentially unique for each product, company, and
industry.

- The process for developing the Summary Function should be generated by an
expert panel, and the process documented. Each factor was defined by factor

specific sub-factors which are covered next.

3 -3 - 5 Proposition 3

Proposition 3 — It is logical to define each factor as a group ofsub-factors, and to define
th 6 root factors by data or reference.

To provide additional insight on the factors, it is logical to break them down to
sub— factors (Figure 6). As was stated in Section 2.1 above, it was not efficient or
efi‘ectiVe to identify common sub-factors that would cover all counterfeit products. The
sub.- faCtors are to be identified and ranked by an expert panel by the users of the model.
To insure process transparency, the derivation of the sub-factor ranks should be

supported by references to reports or data, or at least expert opinion.
W The root factors are broken down into three (or more)
Sub‘ faCtors, to further define the risk. Due to the many potential variables, it was not
efli Ci ent to identify sub-factors that are specific to all products and supply chains. For

S a1lalysis, the formula A+B+D=D considers the risk rank for Sub-Factor 5.1 (a), Sub-

109

 

Factor 5.2 (b), and Sub-factor 5.3 (c) to determine the Total Rank for the Factor 5.
Market Profile. In this example, the Total Rank was “Medium.”
The following support the Concept Statement.
a. Breaking the concept down to smaller and more precise pieces leads to a greater
understanding of the underlying drivers.
0 Due to the nature of specific products, categories and distribution systems, it was
not efficient to standardize sub-factors for the overall Counterfeit Product Risk
Model. This states that it is valuable to see the detail of a completed CPRM
analysis but not to standardize the sub-factors in the general model.
0 Published literature and reports (Chapter 1) recommend that expert panels be
assembled to identify the pertinent sub-factors.
- This process should be simple, efficient, and effective in the hazard identification

stage of risk assessment.

Factor 5. Market Profile

 

¥ Very High High Medium Low Very Low
I Total Rank X

¥

 

 

 

 

 

 

 

 

 

 

‘

Summary Function (for example) I A+B+C=D

 

 

 
  
  
 

S - 1 Contract Manufacturing X
5 - 2 S ingle Distributors per Country X
5 - 3 Re-Furbished Market X

Fis ub‘factors list for example)
gun-e 6- Example of the Sub-Factor Derivation of the Market Profile Factor

 

 

 

 

 

 

 

 

 

Concept Statement 3-2: It was recommended that the sub-factor ratings be
d -
Lved through a quantitative matrix (Figure 7). Each number will be referenced to a

(lat
a Set or at least to expert opinion of how the rank was established. This makes the

110

 

h 32 ard identification stage of risk assessment process efficient and effective. These
points support the Concept Statement.

- To develop a rank of the sub-factors, a data table was created.

«- Due to the nature of specific products, categories and distribution systems, it was
not efficient to standardize this content for all products. This states again that it is
valuable to see the detail of a completed CPRM analysis but not to standardize the
sub-factors in the general model.

- It was recommended that an expert panel be assembled to identify the sub-factors

and to derive the ranking of the sub-factors.

Factor 5. Marketing/ Sub-Factor 5.3 - Re-Furbished Market

 

Very High High Medium Low Very Low
LTotal Rank X

 

 

 

 

 

 

 

 

 

S um mary Function (for example) =B

 

 

Rank Reference/ comment

 

 

 

 

 

 

 

 

Very High 1000 Units Per report A
High 100 Units Per report A
Probability Medium 10 Units X Per report A
Low 1 Unit Per report A
Very Low 0.1 Unit Per report A

 

 

 

 

Figure 7. Example of Sub-Factor Derivation, with Data of the Market Profile Subfactor 5.3

An expert panel should be used to complete the spreadsheet that is shown (Figure

7) illcluded, and to feed (also referred to as a “roll-up” or “cascade” (Howick, Eden et al.,
200 8 )) the results up to the previous spreadsheets (Figure 5 and Figure 6) to build to the

overall Counterfeit Product Risk Model. For this analysis, the formula a=b considers the

ri Sk rank for the derivation of Sub-factor 5.3 (c), and directly enters that rank as the Total

Ra h
k for the Sub-Factor 5.3. In this example, the Total Rank is “Medium.”

111

 

3-4 Model Development Summary

This Chapter covered the Counterfeit Product Risk Model development, including
a scri es of propositions that helped shape each step of the development. The root
concepts of anti-counterfeit strategy, risk analysis and risk models were covered in the

prcvi ous Research Analysis Chapter.

112

 

4 - (l EXPERT SURVEY

4 - 1 Methods — Expert Survey

An expert panel was presented with a survey to obtain feedback and insight on the

Counterfeit Product Risk Model. The expert panel was selected to represent a wide range
of industries, products, functional groups, company sizes, and levels in the organizations.

The Expert Panel Survey was conducted in one of three ways, face-to-face, on a
conference call, or with the rater reading the PowerPoint and answering the Word
document survey. In each survey, a five-slide executive summary PowerPoint was

presented to the subjects prior to the survey being answered.
The expert panel was selected specifically to provide broad-industry and high-

level executive insight on anti-counterfeit strategy and risk assessment. A sample size of

thirty (30) was proposed (thirty-three were actually used) since this number would not
allow a few surveys to skew the overall results (C-STAT, 2008b).

The propositions were originally developed through a review of literature and

With peer consultations. The survey was designed to insure that the propositions of the

model were logical.
The survey utilized bivariate (yes/no, agree/disagree) responses. A scaling

method such as a Lippert Scale (e.g. rate agreement from “1 strongly agree to 5 strongly

d' . .
1S'f‘lgl‘ee”) was not used srnce the survey intent was to define only strong agreement or

s
tl‘ong disagreement. By forcing strong agreement or strong disagreement, the intent of

the
S‘ll‘vey was to motivate more extensive comments. The comments by the diverse and

113

 

expert panel would be very valuable to the nuance of the propositions and of counterfeit

risk assessment.
The use of “it is logical to...” in the survey questions was selected to even further

challenge the Expert Panel with such a definitive phrase. As with choosing the extreme
bival'iate responses versus a Lippert Scale, the intent of the survey was to motivate more
extensive comments. A less definitive description, such as “acceptable,” may not have
forced the high-involvement and self-assessed expert panelists to state a certain position

of strong agreement or strong disagreement. Since the group consisted of senior

personnel who are very involved in risk or anti-counterfeit strategy (as is evidenced in the
assessment of expertise question summarized in Table 16) there was no apprehension that
tlle expert survey participants would be influenced to agreeing with “it is logical”.

This section presents the survey (Figure 8).

114

 

urvey Question 1 — Subject Classification

8) Title: C / VP / Dir / Mgr
Industry: Food / Pharmaceutical / Financial / Ins / Law / Retail / Assn / Academic/ Other

b)
S ize Revenue: +353 / 1-513 / lOOM-lB / lO-IOOM / -lOM

   

SurVey Question 2 — Subject Expertise
Please rank your level of familiarity or expertise in these areas on a scale of 1—5, with 1 being low

 

 

.
and 5 being high. N/A can be used for not applicable.
- Each survey member was selected for specific expertise in areas that contribute to the research.
. FOCUS AREA (SURVEYOR SELF-ASSESSMENT) (5 is high)
a) RJsk, Risk Assessment, and Risk Models (1,2,3,4,5)
1;) Insurance, Financial, or Legal Risk Management (1,2,3,4,5)
c) Food, Product or Asset Brand or Packaging Management (1 ,2,3,4,5)
d) Food, Product or Asset Protection (1,2,3 ,4,5)
6) Food, Product, or Asset Law (1,2,3 ,4,5)
f) Corporate Security (1,2,3,4,5)
g) Anti-Counterfeit Strategy and/ or Risk Models (1 ,2,3,4,5)
Survey Question 3 - Proposition 1 - Assess Type of Model
a) The probability versus severity matrix, using qualitative ranks, and focusing on the probability
portion is a logical starting point for the Counterfeit Product Risk Model. (agree, disagree,
comments)
b) Are there any other references or examples that would support this “assess the type of model”
development?
c) Are there any additional comments, thoughts, suggestions, or questions that we should consider?

 

Survey Question 4 — Proposition 2a — MECE Factors
It is logical to break down the overall Counterfeit Product Risk Model to a mutually-exclusive,

 

 

 

    
    
    
    
 

3)
. comprehensively exhaustive (MECE) set of factors. (agree, disagree, comments)
b) Are there any other references or examples that would support this development?
C) Are there any additional comments, droughts, suggestions, or questions that we should consider
Survey Question 5 — Proposition 2b - Specific Factors
a) For the factors, it is logical to use: Counterfeit-History, Counterfeit-Ability, Counterfeit-
Attractiveness, Counterfeit-Hurdles, and Market Profile. (agree, disagree, comments)
1") Are there any other references or examples that would support this development?
C) Are there any additional comments, thoughts, suggestions, or questions that we should consider
\
Survey Question 6 — Proposition 2c — Summary Function
a) It_ is logical to use a mathematical summary function to summarize the risk to one rank. (agree,
b) d lsagree, comments)
c) Are there any other references or examples that would support this development?
Are there any additional comments, thoughts, suggestions, or questions that we should consider

 

  

S:- h ,
a) e): Question 7 - Proposition 3 — Sub-Factors
It IS logical to define each factor in light of sub-factors, and to define the root factors by data or

b) reference comments. (agree, disagree, comments)
.3) Are there any other references or examples that would support this “model building” development?
Are there any additional comments, thoughts, suggestions, or questions that we should consider?

 

P y Question 8 — Discussion
1eElse comment on the discussion points (agree, disagree, comments)
1 \re there any other references or examples that would support this “type of model” development?

<3)
‘ \1 e there any additional comments, thoughts, suggestions, or questions that we should consider?

   
 
 
 

 

Fi
gure 8. The Expert Panel Survey Questions Based on the Propositions

115

 

4- 1 -1 Statistical Analysis of the Survey Data

Survey question one and question two included demographic information that was
co 1 1 ected from the anonymous survey participants. The information was gathered to
assess the applicable insight of the survey participants. Survey question one included the
demographic information gathered in the survey example above, and included title,
industry, and company revenue. Survey question two included a self-assessment of
expertise in areas that are applicable to risk, the counterfeit threat, corporate security,
1 aw, or packaging. These two questions were gathered for background information and
were not statistically analyzed.
The statistical analysis utilized an “assessing agreement” concept (C-STAT,
2 00 8b). Since the data was in high agreement, it was appropriate to review all the
responses as a total number rather than by each of the six data questions. If the data had
been in less agreement it would have been more appropriate to analyze each question
individuany (Fitzmaurice, 2002b; Sim & Wright, 2005).
To review the overall agreement of the surveys with the research propositions,
two null hypotheses were proposed and the results are presented in Chapter 3.
- Null Hypothesis One: Agreement with the propositions - For the survey participants,
tl'lere is no positive agreement with the propositions.
- Null Hypothesis Two: Agreement more than random - For the survey participants,
there is no more agreement than might occur by random chance.
4 ‘ 1 -2 Null Hypothesis One — Agreement with the PrOpositions
To address Null Hypothesis One, the “observed proportion of overall agreement”,

01‘ ‘ c
positive agreement,” method was utilized (C-STAT, 2008b; Fitzmaurice, 2002a,

116

 

2002b; Spitzer & Fleiss, 1974). In this method the overall positive proportion (“Po+” or
e‘Ppos”) was given by the following expression (Equation 2) which is adapted to the data
chture of this research and is given by the following expression (Fitzmamice, 2002b),

PM = A / N*n (2)

Where “A” is the total responses in agreement with the question, “N” is the
number of questions and “n” is the number of raters. The result is a numerical expression
flex!) 0 to 1. The interpretation of agreement utilizes the information in Table 12 and was
developed through the work of several researchers before assembly into the table and
does not include a “perfect agreement” level (Cohen, 1960; Gwet, 2001; Sim & Wright,

2005). The final table is from Sim & Wright (2005).

Table 12. Interpretation of Agreement for Expert Panel Survey Results

 

 

 

 

 

 

 

Kappa Interpretation
<0 No Agreement
0.0 to 0.20 Slight Agreement
0.21 to 0.40 Fair Agreement
0.41 to 0.60 Moderate Agreement
0.61 to 0.80 Substantial Agreement
0.81 to 1.00 Almost Perfect Agreement

 

 

 

 

This method uses the kappa function (addressed below) and does not traditionally
use Confidence intervals, standard deviations or other statistical analysis (Cohen, 1960;
E‘?il‘:21'llaurice, 2002b; Spitzer & Fleiss, 1974). Classical statistical methods do not apply to
is SeSsing agreement since the research results are agree/ disagree/ comment. While the

an .
SV"V€:rs could be converted to numbers there is no logical way to determine an analytical

117

 

average or a deviation (Capra & Canale, 1998). Also, there is no normal distribution

associated with the dichotomous (yes/no) results.

Since the survey results were in such close agreement, the total responses for the
questions will be reviewed in “total” rather than by individual question.
4 - 1 .3 Null Hypothesis Two - Agreement more than Random
To address Null Hypothesis Two, the “Fleiss’-Kappa” method of assessing
agreement was utilized (C-STAT, 2008b; Cohen, 1960; Fitzmaurice, 2002b; Fleiss, 1971;
F l eiss & Cohen, 1973). The statistical concept of “kappa” is an index of expected
ageement due to chance alone (Cohen, 1960; Fitzmaurice, 2002b).
The Fleiss’-Kappa assessing agreement measure is designed for categorical rating
(e- g - identifying a specific illness from a group or a bivariate response such as yes/no and
not for a range such as a Lippert Scale such as a “1-5 rating” continuum of the magnitude
of agreement) agreement among several raters, whereas other methods, such as Cohen’s—

I(ii-IDIDa, are traditionally focused on comparing agreement between two raters.

Kappa is given by the following expression (Equation 3) (Fitzmaurice, 2002b),
“k”= Po - Pc/ 1 - Pc (3)
The “k” factor is kappa. P0 is the proportion of overall agreement. Pc is the
proportion of overall agreement due to chance. The denominator, 1 — Pc, is the
2:11 ax iInum agreement above chance and the results are presented in a numerical

expr - 99 ° 39 cc
essron fiom -1 to 0 to 1 (from complete disagreement, to no excess agreement

b
eand that of chance” to “perfect agreement”) (Fitzmaurice, 2002b).

118

 

4-2 Expert Survey Results and Discussion

The Expert Panel Survey was conducted to gain insight and feedback on the

C ounterfeit Product Risk Model. The expert panel was selected to provide a broad and
tl—lorough review of the packaging, product protection, counterfeit, and risk disciplines.
The companies or organizations that were included in the panel are involved in doing risk
assessment and/or developing anti-counterfeit strategy. The companies were identified
fiom industry and academic suggestions, and the final selection was based on providing a
wide-range of perspectives for both risk assessment and anti-counterfeit strategy.

The individual panelists were selected by their companies or organizations due to
their applicable insight and expertise. In some cases the anti-counterfeit strategy subject
m atter expert was identified from the outset. In other cases, the researcher worked with
Several company representatives to identify and recruit the most appropriate panelist. In
eaCh case, the subject was approved to participate in the survey. The requirement for
Se] ection as a panelist was determined by balancing a broad and deep perspective on the
Counterfeit Product Risk Model. This perspective was achieved, as evidenced by the

Summary of the panelists in Table 13 (Title), Table 14 (Industry), Table 15 (Company

I{e"enue), and Table 16 (Expertise).

The companies or organizations that were included in the panel are involved in

Olng risk assessment and/or developing anti-counterfeit strategy. The mdrvrdual

D an elists were selected by their companies or organizations due to their applicable insight

3‘6 expertise.

119

The research goal was to validate the counterfeit risk, to review existing models

or concepts, to confirm a series of propositions that will assist future researchers, and to

conduct statistical analysis on the survey inputs.

4 -2.1 Subject Classification Responses — Title,

Industry, Revenue, and Subject Expertise

Regarding the “Subject Classification — Titles”, the inclusion of high-level
executives who are aware of, and versed in their company’s enterprise risk assessment

and anti-counterfeit strategy, was deemed to be especially important (Table 13).

 

 

 

 

 

 

 

 

 

 

 

 

Table 13. Subject Classification — Title of the Expert Survey Subje[0]cta
S uney Question 1 — Subject Classification
2.. Table: C/ VP/ Dir/ Mgr Number
Chief-Level-Officer/ Partner 6
V ice-President 1 l
imam lO
met/Engineer 6
HAL 33

Regarding the “Subject Classification — Industry,” industry coverage including

finarlce, insurance, law, risk theory, and a range of consumer products (Table 14).

120

Table 14. Subject Classification - Industry of the Expert Survey Subjelolcts

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/——S—:nirvey Question 1 - Subject Classification
1). Industry: Food / Pharmaceutical / Financial / Ins I Law / Retail I
Assn / Academic! Other Number
FFood/Beverage 6
Pljarmaceutical/ Healthcare/ Medical Device 8
F inancial 1
Insurance 1
Law 2
Retail 3
Association 2
Academic 1
Other 9
Other-Supplier 1
Other-Consumer Packaged Goods 1
Other-Original Equipment Manufacturer 4
Other-Consumer Electronics 2
Other-Software 1
TOTAL 33
.\ _

 

 

 

 

Regarding the “Subject Classification — Company Size”, it was especially
i1171130rtant that the company size coverage include large (Fortune 500) corporations who
haditionally have more management resources to apply to more precise portions of their
business. The range of companies also included smaller companies that added value to
the overall perspective of the research due to their market niche or expertise. In some
cages, such as specialty law firms, the company revenue is relatively lower, compared to

I 3 roduct manufacturers.

121

Ta ble 15. Survey Classification - Company Revenue of the Expert Survey Subje[0]cts

 

 

 

 

 

 

 

 

 

F S nrvey Question 1 — Subject Classification

c - Company Revenue: +353 (5000)/ l-SB (1000) / 100M-1B (100)/ 10-

1 00M (10)/ -10M (1)] NA Number
>558 (Fortune 500) 16
$ 1 —SB (Fortune 500-1500) 6
S 1 COM-$1B 5
s l OM-$100M 3
<$ 1 0M 0
NA (Assn or Academic) 3
TOTAL 33

 

 

 

 

This section covers the results and discussion of the Subject Expertise questions.

Expert Status is defined as the raters’ self-assessment on a scale of 1 to 5 (5 being high)

(Table 16).

T‘ble 16. Results of Survey Question 2 - Subject Expertise (Self-Assessment) for the Expert Survey

 

 

 

 

 

 

 

 

 

S ‘- bjcflllcts
Average
Survey Question 2 — Subject Expertise (Ranked l,2,3,4,5 with 5 Self- Expert (4 or 5)
m Highest) Assessment Self-Assessed
Us, Risk Assessment, and Risk Models (1 ,2,3,4,5) 3.5 17
.b\.lnsurance, Financial, or Legal Risk Management(l,2,3,4,5) 2.6 9
c ‘ Food, Product or Asset Brand or Packaging Management
&3AS) 3‘7 21
Law, Product or Asset Protection (1,2,3,4,5) 3.6 22
New, Product, or Asset Law (I ,2,3,4,5) 2.7 8
t‘~\(zorpomtc Security (1,2,3 ,4,5) 3 ll
mun-Counterfeit Strategy and/ or Risk Models (1,2,3,4,5) 3.1 11

 

 

 

At least one-third of the raters expressed expert status for each question. It was

Qt surprismg that the average Subject Expertlse rating was moderate (around three) smce

“3 Wide range of industries represented by the raters would not lead to high expertise in

122

 

al 1 areas. Due to the make-up of the industry representation, it was not surprising that
brand, packaging and protection areas had so many experts and a high average
as sessment. With anti-counterfeit strategy being a relatively newer concept, it was

somewhat surprising that there was such high average expertise.

4 - 2.2 Statistical Analysis - Propositions, Survey

Results

The statistical analysis of the survey is reviewed here. The survey raters

responded with agree, disagree, comment, or no response. Since “no response” or

‘

‘comment” neither signified agreement nor disagreement, these last two responses were
considered in a separate, single, third category. For assessing agreement, only “Agree”
was used for the affirmative response.

The survey results emphasize that the Counterfeit Product Risk Model is logical

and practical for industry and agencies. The survey responses are summarized in Table
1 : -

123

Ta ble 17. Survey Results of the Expert Survey Subje[0]cts

 

 

 

 

 

 

 

 

{t No
Answer/

Question (N =33) Agree Comment Disagree
S urvey Question 3 - Proposition 1 —- Assess Type of Model 31 1 1
Survey Question 4 — Proposition 2a — MECE Factors 27 4 2
Survey Question 5 - Proposition 2b - Specific Factors 33 0 O
S urvey Question 6 - Proposition 2c — Summary Function 31 2 0
Survey Question 7 - Proposition 3 - Sub-Factors 32 l 0
Survey Question 8 — Discussion 32 1 0
TOTAL 186 9 3

 

 

 

 

 

 

4 -2.3 Null Hypothesis One - Agreement with the Propositions

N 2411 Hypothesis One: Agreement with the propositions - For the survey participants,
th ere is no positive agreement with the propositions.
To address Null Hypothesis One, the “observed proportion of overall agreement”
In ethod was utilized (Fitzmaurice, 2002b; Spitzer & Fleiss, 1974). This process,
di Scussed in the Methods (presented in Section ## to ##), assesses the overall agreement
betWeen many raters on dichotomous (agree/disagree) type surveys. For the thirty-three
SI"~1‘\’eyors, the “observed proportion of overall agreement,” or “positive agreement,” is

0 r 93 which is interpreted as “almost perfect agreement (C-STAT, 2008a). The results

124

reject Null Hypothesis One and supports that the survey subjects agree with the
Counterfeit Product Risk Model, and that the results can be interpreted as an “almost

perfect agreement.”

4.2.3.1 Null Hypothesis Two - Agreement more than Random

Null Hypothesis Two: Agreement more than random - For the survey participants, there
is no more agreement than might occur by random chance.

To address Null Hypothesis Two, the “Fleiss-Kappa” method was utilized
(Cohen, 1960; Fleiss & Cohen, 1973). This method (presented in section ## to #)
assesses the agreement more than on a random basis of raters for dichotomous
(agree/disagree) type surveys.

For the thirty-three surveys, the “F1eiss’-Kappa” number was 0.90, which is
interpreted as “almost perfect agreement” (C-STAT, 2008a).

This result rejects Null Hypothesis Two and supports the concept that the
agreement is more than what would be expected by random chance, and the results are
interpreted as an “almost perfect agreement” between the survey subjects.
4.2.3.2 Summary

Utilizing “assessment agreement” statistical analysis, the positive agreement for
the thirty-three surveys was 0.93 or an “almost perfect agreement” and the Fleiss’-Kappa
was 0.90 or an “almost perfect agreement.” Thus, the results indicate that the survey

raters statistically agree with the propositions and the Counterfeit Product Risk Model.

The fill] raw comments are included in the Appendix C.

125

4.2.4 Comments for Question Three — Assess Type of

Model

The comments for question three supported using the probability versus severity
matrix, qualitative factors, and focusing on the probability portion of the probability
versus severity matrix. An interesting insight was that some companies do “baseline”
anti-counterfeit (e. g. holograms) or product protection (e. g. tamper evidence) actions for
all products. Any anti-counterfeit or product protection initiatives above that “baseline”
would need to be justified, compared with all activities (operations, marketing,
advertising, new product development, etc.) across a brand or company, considering the
return-on-investment.

The model can be used to compare counterfeit risk across products, whether they
are known counterfeited products or not, or even if they are products that have not yet
come to market. Specifically, there was a comment that the Counterfeit Product Risk
Model could be used as a preventative measure, to evaluate food economic adulteration
risks.

The comments emphasized the importance of providing a theoretical justification
for got using quantitative processes such as FMEA. The theoretical justification will help
with some companies specifying the use of the traditional risk or quality assessment
tools.

Regarding the emphasis on qualitative ranks, one comment was that “[It] is also

IOgical to use qualitative descriptions for the final ranks since the 'data' is based on

assumption and not actual tests, which you can do for F MEA.”

126

The comments support the methodology used and standardization employed in the
Counterfeit Product Risk Model. Not going directly to a quantitative model is helpful, in
that it is important to clearly define the factors and the derivation of each factor. That the

model and process need to be continually updated for emerging threats.

4.2.4.1 Comments for Question Four - MECE Factors

The comments supported using mutually-exclusive and comprehensively-
exhaustive (MECE) factors, though there was concern that too much rigidity in the
process or model could cause emerging risks to be overlooked. One anti-counterfeit
expert rater noted “It is very hard to force the many aspects of the counterfeit threat into
finite categories, and you may miss some of the real issues.” Another anti-counterfeit
strategy expert rater noted “[MECE] is difficult to accomplish, however, since factors
will often intersect and have overlapping cause and effect relationships. On an informal
assessment of risk factors, what seems illogical to the brand owner may in fact be a good
business proposition to the counterfeiter.”

These comments support the methodology and standardization in the Counterfeit
Product Risk Model. A continuing theme of the comments was the importance of clearly
defining the factor definitions and the derivation of the factors.
4.2.4.2 Comments for Question Five - Specific Factors

The comments supported using the specific factors of Counterfeit-History,
Counterfeit-Ability, Counterfeit-Attractiveness, Counterfeit-Hurdles, and Market Profile.
These comments further support the methodology and standardization in the Counterfeit

Product Risk Model. The comments noted a continuing theme which was to clearly state

the factor definitions as well as the derivation of the factors.

127

For example, “From years of experience in anti-counterfeit, these five categories
cover the range of counterfeit and diversion factors. I have not seen such a detailed and
complete list of factors. Excellent.”

There was the same concern as for the MECE question, regarding losing some
model sensitivity by forcing characteristics into specific factors. Also, there were
comments that these five factors could be combined into fewer categories, such as
“history,” “ease of entry,” and “profitability.” More, and fewer, categories were
considered, but the combination of these five factors and sub-factors is simple and yet
detailed. Adding a layer of hierarchy would complicate the practical implementation of
the model.

The use of an expert panel to specifically define the factors and the derivation of
the factors reduces concerns that companies or agencies might have. Also, the factors are
defined by the sub-factors, so any specific area of concern would be defined when
assigning the sub-factors. For example, if the level of contract manufacturing is a
concern, that specific characteristic can be placed into a sub-factor in “Counterfeit-
Hurdles” and then the sub-factor helps define the factor. The comments supported the

methodology and standardization in the Counterfeit Product Risk Model.
4.2.4.3 ~ Comments for Question Six - Summary Function

The comments further support the methodology development and standardization
of the Counterfeit Product Risk Model, specifically using an expert panel derived
mathematical formula to decide on the “Total Rank” of the overall probability, the factors
and the sub-factors. There were a few responses that did not “agree” with the question,

due to concerns with how the mathematical formulas would be derived. Also, there was

128

concern that the formulas (as well as the factors or sub-factors) could become inflexible
and hinder the value of the model.

Some raters felt that it was important to not force the results to numbers too early.
The rater was comfortable that numbers were used for the summation, but that qualitative
results are used in the matrix.

This will not be an issue since general risk assessment methodology recommends
continually monitoring the changing risks. Also, the model building process emphasizes
using an expert panel to make sure the model works for the specific product or company.

One comment directly addressed the findings hour the literature review, that the
traditional operational or quality control process, FMEA simply cannot be accurately
applied to the counterfeit risk.

Again, the use of the expert panel was supported by the following statements in
the survey comments “Using expert panel is the way to look at this risk, we attempt to
use some modeling plus SME (subject matter expert) insight. Mostly SME [subject
matter expert, and official designation in some companies or agencies].”

There was a continuing concern with quantitative measures possibly obscuring
risks. “1 think the mathematical summary would directionally predict the potential risk,
but it still leaves a chance that a lower ranked risk could be a problem.” The comments
further support the methodology and standardization in the Counterfeit Product Risk

Model.
4.2.4.4 Comments for Question Seven — Sub-Factors

The comments supported the methodology and standardization in the Counterfeit

Product Risk Model, and specifically the use of sub-factors and of clearly documenting

129

the derivation of those sub-factors. The comments clearly supported not defining
common sub-factors for all products or industries. “These categories can be quite
squishy if you do not clearly define them. Also a straight interpretation for all industries
or all products is very difficult and probably dangerous.” “There are just so many
subjective breakdowns you can make before they become bias. It is not clear you need
sub-factors.”

Another rater noted that the extra step of using sub-factors was novel and helpful.
“It is excellent to define the factors by the sub-factors. We have not gone to that detail
and this is very helpful when discussing the details of the prediction with decision
makers.”

The comments further support the methodology and standardization in the
Counterfeit Product Risk Model. Specifically the raters embraced the concept of further
refining of the factors to sub-factors. It was also noted that referencing specific reports or
methods of deriving the sub-factors would give credibility to the overall ranking of the
factors and the overall counterfeit risk probability.
4.2.4.5 Comments for Question Eight — Discussion Section

The comments further support the methodology and standardization in the
Counterfeit Product Risk Model. The comments included: “I believe you need to follow
a proven methodology as a beginning point then maybe tailor the survey/tool or
instrument to a specific industry, product, etc.” And specifically in support of the
scholarly overview and theoretical defense of the model: “I am constantly challenged by

no academic work in this counterfeit risk area - that is challenged by managers asking for

130

methods and for my trying to develop the tools” and “There are very generally accepted
attempts at the analysis from other disciplines, but no theoretical justification.”

The comments further support the methodology and standardization in the
Counterfeit Product Risk Model. The expert panelists agreed with the general and the

specific steps presented.
4.2.5 Expert Survey Summary

The survey resulted in “almost perfect agreement” with the propositions for the
Counterfeit Product Risk Model using a very representative expert panel. The raw

comments provided excellent insight and support of the propositions.

131

5.0 CASE STUDY

A food ingredient company (referred to as “CSFI”) was confidentially engaged to
conduct this case study to demonstrate the use of the Counterfeit Risk Model (Model). A
series of interviews were conducted to solicit background information, to gain insight
from a company expert panel, and to discuss both the working of the model and the
results.

The company manufactures branded food ingredient products in a single-use
“form” combined in a multiple-unit primary package selling unit. The product is sold by
business-to-business (B2B), business-to-consumer (BZC), and also there is evidence of
selling by consumers-to-consumers (C2C) on websites such as “e-Bay.” The privately
held company was started between 1985 and 1995 and has revenue in the $5M-$20M
range. The company has global distribution of their product in both the bulk and finished

form, and manufactures 95% of their product in one, US based facility.

5.1 Methods

The researcher visited the company and conducted follow-up interviews with key
personnel to gather background and specific information on the CSF I operations
(procurement, manufacturing, distribution, customer service, and others) and the
marketing (competitive products, sales process, distribution, retailing, and others). The
researcher and the key company contact (the Vice-President responsible for brand
integrity) developed the preliminary CFSI data that would be entered in the Counterfeit

Risk Model (Model).

132

The researcher and key contact developed the first versions of the completed
Model which were presented to the company expert panel (the full leadership team) and
used this larger corporate group as the gathered expert panel. The expert panel discussed
the sub-factors, sub-factor derivation, sub-factors that were not included at this time (and
which would be used for future possible review), the summary formulas, and the overall
results of the Counterfeit Risk Model.

The expert panel chose to narrow the details of the product, geography, and

channel.

0 Product: A specific CSFI branded, packaged product that is sold by anyone
directly to consumers (B2B, BZC, or C2C), including all forms of the product and
all package styles including promotional items or special packs.

0 Market Geography: The specific geography was for sales in the US.

0 Market Channel: The specific market channel was actually all channels including

wholesale, retail, on-line retail, and auction sites (e-Bay, etc). .

5.2 Results

The Model was used with CFSI data and presented in an overall counterfeit risk in
Table 18 (the table is at the end of this section and the full analysis is in Appendix E). To
protect the identity of the company, the sub—factor derivation descriptors (identifying why
a sub-factor was very high, high, medium, low, or very low) was recorded as “tbd” (to-
be-determined) and the sub—factor derivation comments were modified to protect the
confidentiality of the company.

The result of the Counterfeit Risk Model was that CSF I has a “Low” probability

of counterfeit risk because: (1) 95+% of product is manufactured at one proprietary

133

location on proprietary equipment, (2) the company has comparably low volume

combined with advanced packaging and product features that are somewhat complex to

duplicate, and (3) the company monitors product in the marketplace including online

C2C.

Regarding the Factors and Sub-Factors, it should be emphasized that it is up to the
expert panel to define how sub-factors are arranged and weighted. Unique production,
supply chain, marketing, retailing, or consumer factors may not be apparent to someone
outside the company.

Details of the “Factors” are included here.

0 Counterfeit-History. This Total Rank was “low” since there are no known counterfeit
products, there is little evidence of diversion of product outside the legitimate supply
chains but there is some risk since related, lower premium priced products, are known
to be counterfeited and diverted. These points lead to a “Low” rank because CSFI is
very diligent in monitoring the market and has no evidence of counterfeit product and
very little diverted product.

0 Counterfeit-Ability. This Total Risk was “Medium” since the basic product form is
fiequently counterfeited but not “High” since CSFI uses very unique and complex
packaging that would increase challenges for counterfeiters. These points lead to the
increased “Medium” rank because even though there is a no evidence of
counterfeiting of this product, there is a high likelihood that counterfeiters have the
capabilities to attack.

0 Counterfeit-Attractiveness. The Total Rank was “Low” even though the product has

a premium price of two to five times more than the generic version of the product.

134

 

Tight supply chain monitoring and supply chain relationships make it hard to place
counterfeit product in the legitimate supply chain. These points lead to a “Low” rank
because even though there is a high profit-margin between the cost of generic goods
and the CSFI product, there are many supply chain actions that make it difficult for
counterfeiters to get, and keep, product in the legitimate supply chain.

0 Counterfeit-Hurdles. The Total Rank was “Low” since there are many functional
packaging features (tamper-evidence, child-resistance, print carton tape, and others),
as well as other supply chain controls. These points lead to a “Low” rank because
these actions lead counterfeiters to need more complex production or packaging, or
they must overcome supply chain controls. (For continuity in the model, “low” is
good, so “Low” in this rating means that there are a many or complex hurdles, which
lead to a “Low” risk.)

0 Market Profile. The Total Rank was “Low” since very brand loyal consumers are
fairly price insensitive and pursue reputable sources of supply, and 95%+ of the
production and warehousing is in-house. These points lead to a “Low” rank because
the consumer does not go looking for good deals from un-reputable sources, and with
in-house production there is little risk of unauthorized production or co-mingling by

contract manufacturers.

5.3 Case Study Summary

The Counterfeit Risk Model was successfully used to review a company’s
Specific counterfeit risk attributes. The process of completing the Counterfeit Risk
Model provided a mechanism to gain corporate consensus and set clear direction for

effective countermeasures in product development, regulatory affairs, customer service,

135

 

 

and legal strategy. The sub-factor derivation discussion and analysis was valuable to
allow reproducing the analysis in the future and to consider how new knowledge of
emerging threats change the overall risk.

Reviewers not involved in the process may perceive that some Factors or Sub-
Factors overlap, but there was agreement by the Expert Panel.

Bringing the expert panel together was useful since it brought issues to light/
discussion in a very focused review of the counterfeit threat, so that the overall company
more clearly understands their actual risks. CFSI considered the process efficient to

conduct the risk assessment and as an internal communication tool.

136

Table 18. Case Study Overall Counterfeit Risk for CFSIloI

Case Study: Confidential Food Company
Bolded Box = Data Entry
1. Overall Counterfeit Risk

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Probability
"= 3
00
E a .5: 3 .3
:9 “a": ‘3 3 s
> 2 :>
(1) 95+% of product is manufactured at
one proprietary location, (2) The
company has comparably low volume
R 1 combined with advanced packaging and
Total X product features that are somewhat
complex to duplicate, (3) the company
monitors product in the marketplace
includingflrline C2C
J actors
Summary Function: A +B+C+D+E=F [ 1 All equal, manual derivation
¥ No known counterfeits, little diversion,
1 . Counterfeit-History 0 0 0 X 0 but related lower premium product are
faked
The product and from are frequently
2 - Counterfeit-Ability 0 0 X 0 0 counterfeited, but unique feature reduce
.‘ risk
. This is a premium product with $2-5x
3 - Counterfert- . . . . .
Attractiveness 0 0 0 X 0 generlc pncrng, but tight supply chain
‘ controls
High hurdles for counterfeiters, in terms
4 . Counterfeit-Hurdles 0 0 0 X 0 of pkg features, tight supply chain, and
\ monitoring
Brand loyal consumers, price insensitive,
5 . Market Profile 0 0 0 X 0 pursue reputable sources, and in-house
\ production

 

 

 

 

 

 

 

 

137

 

 

6.0 CONCLUSION

Companies and agencies constantly struggle to quantify the magnitude of the
counterfeit threat from both a global and a specific product perspective. Although there
are many examples of the dangers of product counterfeiting, the nature of the
counterfeiters and counterfeiting makes it difficult or even impossible to determine the
quantitative hard data on the risk. Specifically, analysis of the risk, risk model literature
review, in combination with peer-consultation, established a foundation for the
Counterfeit Product Risk Model and for the supporting non-quantitative analysis. It is

not practical to conduct quantitative or classical statistical tool-based risk assessments for
the counterfeit threat because the results can not be theoretically supported. This
research set out to break new ground by presenting an overview of the product
counterfeiting threat as a starting point for the development of a practical, useful and
publically available, Counterfeit Product Risk Model.

This research used a very broad, very interdisciplinary perspective that led to
inlportant theoretical justifications, such as using a probability versus a severity matrix,
qualitative ranking, and the language of enterprise risk management. Many current anti-
<=<>unterfeit research projects are extensions of existing research (with a narrow focus that
is not all-encompassing) or are so theoretical in nature that theyare not applicable (e.g.
Very complex models that are not all-encompassing). This analysis and the development

of the model, provides a unique and practical approach in implementation of anti-

counterfeit strategies.

138

This type of research analysis and model have not been presented previously by
other researchers for several reasons: the extremely interdisciplinary nature of the
strategy; the fact that the hard data is elusive or non-existent—current industry actions
are usually confidential and agency actions are usually classified; broad risk assessment
is still evolving; and a risk-based approach to regulation and legislation is only beginning
to be applied to risks that are very real but very qualitative and hard to evaluate.
Packaging is a logical starting point for this anti-counterfeit research, since the most
efficient anti-counterfeit actions are packaging components, but there are many other

disciplines that are equally important in an anti-counterfeit strategy. Critical disciplines
Which should be considered in an all-encompassing, strategic perspective on deterring
counterfeiting, include Criminal Justice, Supply Chain, Risk Management, Social
Anthropology, Consumer Behavior, Health Risk Communication, Retailing, Intellectual
Property Rights Law, Food Law, Health Care (Medicine, Nursing, etc.), Public Health,
Political Science, International Trade Relations, and many more. Since this is probably
no more than ten percent of any one academic discipline, it is not surprising that there is a
lack of research focus and leadership in the area of anti-counterfeit strategy.

It was important that this research stepped back fi'om current research paths and
reViewed the basic propositions that went into an anti-counterfeit strategy. This then, was
established as the first step in developing a Counterfeit Product Risk Model. It was also
important to review risk analysis and risk models, in general, instead of following the
c1lI‘r-ently established quantitative and mathematical model driven research paths. The

ITIDA and USDA are trying to implement more qualitative risk assessment methods where

appropriate, but tradition and the comfort level of the researchers often forces

139

quantitative outputs. Review of the broad, interdisciplinary research in this area was

important for development of the propositions which led to the development of the

Counterfeit Product Risk Model.

The research analysis and model are intended to be a starting point that can adapt
to emerging insights in the science of risk and lead to a clearer understanding of the
actions of the counterfeiters. Discussions with industry representatives and agencies have
underscored the importance of the basic research performed; as well as to more broadly
validate anti-counterfeit strategy as an academic pursuit. Support for this research has
come from academia, industry, agencies, and associations. A scholarly platform has been

established for future research and for activities in anti-counterfeit strategy.
6.1 For Further Research

The very nature of developing propositions for this research established a base for
future research and model refinement. The logical next step is to use the model to gain
insights, refine usability, and to present procedures for practical implementation. As the
model is used in practice, more detailed risk-based and classical statistical tools could be
uSed to better support anti-counterfeit strategy decisions. Valuable insights could be
gained fiom running the model for various industries and by conducting reviews of inter-
illclustry best practices.

Any future research should be combined with the evolving Enterprise Risk
Management practice and with case studies to both understand and support how financial
aIlti-counterfeit strategic decisions are being made within companies and agencies.

Another important — probably the epicenter of all future anti—counterfeit strategy research

140

— is exploring the behavioral aspects of “the chemistry of the crime” and “the chemistry
of consumer consumption.”
Finally, there should be an ongoing review of both the basic propositions and the

model, itself, with refinements implemented as necessary.

141

APPENDICES

142

APPENDIX A — Results of Expert Survey

 

Question (N =33)

Q3

Q4

Q5

Q6

Q7

Q8

 

Agree

31

27

33

31

32

32

 

No Answer/
Comment

w

A

0

~

 

Disagree

O

 

81118.1129

 

81020.1413

>>

>>

 

81024.1031

 

81023.1357

 

81027.0909

 

81027.1014

 

81105.0904

 

81024.1633

 

81104.1535

 

81104.1442

>>>>>>>>>> N

>>>>>>>

 

81104.1403

U

 

81030.1753

>

 

81105.1643

('5

 

81106.1128

 

81029.0832

 

81118.1318

 

81107.0914

>>>>

 

81113.0221

O

 

81114.1307

>

 

81114.1221

O

 

81114.1821

U

 

81114.1556

 

81114.1503

 

81117.1553

 

81118.1026

 

81114.1405

>>>>>>>>>>>>>>>>>>>>>>>>>> .—

 

81117.0840

U

 

81118.1022

 

81118.1359

>>

>>>>>>>>>>>>>>>>>>>>>>>>>>>>> O

 

81119.1200

O

 

81120.1203

>>>>>>>>>>

 

81120.0916

O

 

81115.0808

 

 

>>>

 

>

 

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> O

>>>>>>>>>>>>>>>>>>>>>>

 

 

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

 

>>>

 

Note: Question 1 and 2 were background information and not survey data.
Note: The Survey Identifier is the year (2008 is “8”), the month (November is “11”) the date (twenty-fourth

is “24”) and afler the decimal place is the time on a 24-hour clock.

143

 

APPENDIX B — Raw Statistical Analysis

Table 19. Raw Statistical Analysis-Inputs and Overall Agreement

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 Inputs 1 1 1 1 1' I
_ _ ,_ 1 1 I I 1 _,
Ache"? 1 .w . 1 _ - I
L... _ . - ...i_--. _v - ‘
L _ ___=_ _ 6 Number of Qiestions __1
1____ _ 1n_=_ _ 33 Number of Raters - 1
i: ______k:=_ _ 3 Nunber of Answer types (Agree Dlsagree NA" or Comment)
1 ;N*n 198 Total Answers 1- i
“ l ' T "' - ‘ *
1-- 1 JI. -_____ -1- _ 1 1
1. Survey Response Data (Noteidata as of 1_1/_2/__2008, fourteen raters) _ '
1 fi_ 1 j ___1 1 Answer'l’lpes (k) 1 _ 1 _ -- _,- 1
1 Questions (N) Agree NA/ None Disagree ___ f 1 1 - _fl 4 -
___________3__ 31 1 1 _ _- 1 ___ _ __ _
' 4 27 4 2 _ , _ _ _ _ __
; 5 33 0 0 .. _)__ 1__ ______ ___ __
o 31 2 0 _j __ ___ __ _ _j
j 7 32 1 o - V r - _# ___ 11
i s 32 1 o _ 1 _ - g _ 1
' Total 186 9 3 - 1 _ _
II. Positive Agreement
A. Observed Proportion of Overall Agreement (Po) _ 1 3* ___"; _: --_-- _— _ _::
Formila= LProportion pos_it_iv_e(t_ota1)/ total answers 1 — Hfi— ::
(Total /9) / N*n
1,— _— 41L ___ Mat—.__“.s— 1—— ————— ———— ........... _. _. _ _ __ -
1:1—:__ 1 7 _ _1 71 - _1 1Anaswer Types (k) __ I- 1 :1“- F _w — _ : _ h :11
1____ glee NA/ None Disagree _ . 1 _ 1
P0 = 0.9394 0.0455 0.0152 Should be 0 to 1 g -
l _- ._ _ --.-.._. _ -...1
LB. Proportion of Agreement due to Chance(Pc) 1 1 1
1 i __‘—- 1T"; - T M— —" _'
1: - Feawlai 1.1/Answer laps JET/E. -1 P 1 i .1. _ ___ i H .1
-.... _ -.' -__ __ ___- . .__-.- -- _ . _ -. - - ._ 1' - - - - -
1 _ _ 1A_ns wer Types___ (k) __ __-_-_.-..- fi -—+-+—

__ ALe NA/ None Disggree __ _ 1 _1 ,
__ Pc = 0.3333 0.3333 0.3333 Should _be Qto 1_ j
1C. Pmponion ma age-.as; (1353.7 'i:1'1—:_ ‘_ 1:: if g _ _ 1 fl ‘ 1;

The percent that agrees with the ___p_ posal, l_n_ CLUE case_Po_l_ _ 1 -
_ _ Should be 0tol -__-._..__._. _ ,1 .. _ _
-__- I
_ Ppos = 0.9394 = "Almost Perfect Agreement

144

Table 20. Raw Statistical Analysis-Inputs and Overall Agreement

I 11. Kappa

The measure of variability in the data, namely adjusting for chance agreement
Should be -1 to 0 to 1

 

 

 

Formula= k = P0 - Pc
1 - Pc
k= Po - Pc
1 - Pc
0.9394 - 0.3333
1 - 0.3333
1 k= 1 0.9091 1 "Almost Perfect Agreement"

 

145

APPENDIX C - Survey Responses — Raw
Comments

Note: Survey questions 1 and 2 were background information and not comments were
collected.

3. Survey Question 3 - Proposition 1 - Assess Type of Model

0 It is good to get your arms around probability since severity is very specific to a company's risk
appetite. For example they may all have a moral high point that they first try to achieve, and then they
look at ROI above that.

0 I continually hear from industry that they are at a loss on how to predict which of their ingredients and
supplies are at risk and what the possible adulterants could be. Additionally whether there are any
methods to detect/test for the possible adulterants.

FDA used this approach.
The risk field is interested in the practical implementation of qualitative models to emerging risks - this
is important work.

0 If probability includes some measure of frequency and if severity includes a measure of consequence,
or if frequency and consequence are otherwise captured. Probability should include consistent
likelihood ranges that will cover the continuum of potential scenarios. Severity should also have well
defined and understood consequence ranges. The matrix must be simple to understand and use.

0 Confiscated goods, reported occurrences

0 Similar matrices are used for workman's comp/ injuries; based on hours worked and number of
injuries.

0 Agree, however most anti-counterfeit programs begin as a result of counterfeit activity, therefore the
risk matrix analysis occurs after the problem has begun.

o Might make sense to also survey actual known occurrences to predict future behaviors - e.g. If
someone found value in a specific area, they may do so again even if that area did not come out of the
model as a high risk category.

0 These emerging risk studies include high/ medium/ low, and I like adding very high and very low as
the far extremes. / Severity is a different thing that is very specific to corporations or even to operating
companies. Also a challenge to even conduct the analysis - if we find one counterfeit product, is that
.0001% or 1%? If it is intern additive vs. a deadly compound, and does the risk impact brand damage,
death, or costs of investigating our entire supply chain. / We find our other risk or safety experts trying
to force FMEA type analysis, but it just doesn't work - it is just not suited for this problem.

0 We look at costs/ risks in three categories: safety, brand equity, and actual final loss - but this is really
severity not probability. It is very hard to create a matrix that mixes each type of loss since we have
ways we usually mitigate each category.

0 Our first screen is brand/fmancial consequence to the organization. Once high, medium and low
financial impact products are determined, then this matrix makes sense.

0 I would be interested to know the frequency of known counterfeit situations in multiple industries to
better understand how probability is determined. // Other suggestion is to include information on black
market potential for food items and what drivers make some products more attractive for
counterfeiting.

o This shows where we would want to protect against and not necessarily where we would be
counterfeited

o It might be useful to factor in references or examples of comparable models that rely on other factors
along with a brief description of how the other models fundamentally differ from the proposed model
(assuming such alternate models exist).

0 makes sense to me to start here, but it is sometimes the least expected areas that are hit

146

4. Survey Question 4 — Proposition 2a - MECE Factors

Overlap causes confusion

It seems like that has to be the first step before other approaches are considered.

It is very hard to force the many aspects of the counterfeit threat into finite categories, and you may
miss some of the real issues.

Effective risk analysis is dependent on risks being defined by MECE factors. This is difficult to
accomplish however since factors will ofien intersect and have overlapping cause and effect
relationships. On an informal assessment of risk factors, what seems illogical to the brand owner may
in fact be a good business proposition to the counterfeiter. (Note: agree)

Health/ wellness indicator models may fit into this category

Sure, but recognize that the counterfeiters do not put much thought into their decisions. It is simply a
factor of product popularity and ease of replication.

Tough to suggest unless all know influences were known, and they rarely are all known. The
comprehensive list of factors can never be entirely exhaustive unless multiple sources are compared
and combined to assure the broadest exposure possible.

The definitions will continue to need to be reviewed since the risks and definitions keep changing as
the counterfeiters adapt.

Of course you'd like to try to have all the factors be very distinct but many overlap or are dependent...
but it is still best to try.

factors often have impact on each other in the real world

It seems logical, but I’d like to see the assumptions used and explanation of factors that were used to
predict counterfeit possibility.

logical, yes, but not necessarily practical

not sure what all the MECE factors are (note: agreed below)

5. Survey Question 5- Proposition 2b- Specific Factors

This would probably be considered under Ability or Attractiveness- but it seems like on could
consider Opportrmity. Could the chemical nature of certain food ingredients be cross linked with other
chemicals that are produced cheaply in other industries that could lend themselves to opportunistic
adulteration? The availability, cost, geographic proximity to production, and other factors could come
into play.

You could also consider shock in your market profile (Note: this would be included in Food Defense/
Carver+Shock models)

Definitions of terms is unknown thus difficult to agree or disagree. Is Counterfeit-Attractiveness a
function of Counterfeit -Hurdles and Market Profile? For a specific product Counterfeit-Hurdles are
high and Market Profile is poor would the Counterfeit-Attractiveness necessarily be low? If so the
factors are not MECE. Another example of the interdependency between the factors, (making them
not MECE), would be the relationship between Counterfeit History and the other factors. Occurrence
of counterfeiting in the past is most likely dependent on counterfeit-ability, counterfeit-attractiveness,
etc.

I am assuming that the items listed above would be buckets for which various elements would be listed
within each. Clear concise definitions would need to be established for each category.// Other models
could be NFPA/ Zurich/ Factory Mutual formats for fire loss

Perfect, covers them all.

These are the factors we consider now.

Same comments as 3.

I think this is a good starting point

From years of experience in anti-counterfeit actions and deterrence, these five categories cover the
range of counterfeit and diversion factors. I have not seen such a detailed and complete list of factors.
Excellent.

An issue is that these are subjective, for example where would covert inks go. But this is defined by
the sub-factors. / It is important to consider the type of action, such as counterfeit (pass as legitimate
product) or knock offs (consumers figure it is fake/ stolen).

in addition to use sales/ revenue as well

147

6.

7.

Ability and attractiveness could be considered together since they are very intertwined. It is interesting
what we see counterfeiters focusing on to decide what to fake, for example the large scale
counterfeiters look for chronic products (products taken continually) not remedial (immediate effect).
Definitely will need expert panel to refine the details.

Agree assuming hurdles include covert technology and historical impact on ability to prosecute.

More information on past counterfeit situations with explanation of the drivers that led to these
situations would be helpful

Can it be boiled down to, or have, major buckets of 1. easy of entry and 2. profitability

Agree - Tylenol was a big national brand and if it was CVS brand aspirin, it would not have gotten the
same attention. . ..therefore I believe Market Profile is an important factor. Do you consider price
points of a product as part of the Market Profile? An item that retails for $0.99/unit is not the same as
something like baby formula that sells for $30/unit.

Survey Question 6 - Proposition 2c - Summary Function

Not familiar with it.

I think this makes sense.

There should be more than one since there are multiple factors (NOTE: originally noted as "disagree"
but there are, so it is actually agreement)

When using FMEA's there is a rating of 0-1- and it is logical to rank in this model - but it is also
logical to use qualitative descriptions for the final ranks since the 'data' is based on assumption and not
actual tests, which you can do for FMEA.

Don't know - would have to see variables and formula

Generally accepted method for risk ranking.

Rank would be effective and useful until the counterfeiters learned the process then used the profile to
reassign/ use the matrix to their advantage. Could also see, depending on the amount of categories and
weight of each, a common score - result should provide insight on how an item is suggested to be
handled... if everything is a high, all efforts would be in vain.

Would be nice, as long as the financial benefit is included.

Using expert panel is the way to look at this risk, we attempt to use some modeling plus SME (subject
matter expert) insight. Mostly SME.

need to capture soft or other tangibles when deciding overall approach

If the weighting is correct and can apply across categories.

It might be useful to factor in references or examples that expose any generally accepted shortcomings
or flaws that may exist with respect to using a mathematical approach, assuming any such references
exist

Partially agree. . ..I think the mathematical surmnary would directionally predict the potential risk, but
it still leaves a chance that a lower ranked risk could be a problem

Survey Question 7 - Proposition 3 — Sub-Factors

These categories can be quite squishy if you do not clearly define them. Also a straight interpretation
for all industries or all products is very difficult and probably dangerous. For example, a company
may actually have more control of a contract manufacturer than of their own facility.

Not familiar with it.

There are just so many subjective breakdowns you can make before they become bias. It is not clear
you need subfactors. (Note: "agree" if the sub-factors are not predetermined for all products)

Agree, depending upon how allocated

Important that the root factors are defined by consistent likelihood ranges that will cover the
continuum of potential scenarios and consequence ranges which are well defined and understood. If
the objective is to arrive at a single numerical value, it will be important to weight the sub factors
appropriately and to identify a common scale to use among the sub factors.

Once comments are added in - there is a level of subjectivity that may cloud the results; the listing of
root factors/ comments should be predefined and consistent/ selectable if applicable but same for each
product.

148

It is excellent to define the factors by the sub-factors. We have not gone to that detail and this is very
helpful when discussing the details of the prediction with decision makers. This process is flexible
enough to address specific corporate initiatives and the changing nature of the risk.

It would be interesting to see a food safety risk profile versus our prescription pharmaceutical profile...
or other product for that matter.

Generally agree, but need more information to better understand the relationship of sub—factors and
how this was determined

8. Survey Question 8 — Discussion

This could definitely be used in the HACCP risk assessment step, and most companies do not really to
any emerging risk assessment whatsoever. Your company needs to decide what level of risk it is
willing to accept, and not thinking about it is a decision. Also, it is kind of a psychological thing
thinking that these things are such low probability, but the high severity is where you really get into
trouble.

I believe you need to follow a proven methodology as a beginning point then maybe tailor the
survey/tool or instrument to a specific industry, product, etc. (Note: strongly supports this
methodology)

Counterfeiting risk depends upon potential margin mark-up of counterfeited goods. Thus, low priced
items with large potential mark-up (e.g. Clorox @100% markup and easy preparation/ marketing) is
more attractive than Bulova watches w/ 40% markup and difficult logistics.

In general there is agreement that the methodology and factors presented are a logical way to assess
counterfeit risk. If additional detail were provided to describe the specific factors and examples of
subfactors were included to illustrate the evaluation process, this would be of benefit in the analysis.
Because of the dynamics of economics/ developing technologies etc... the system built should be
dynamic enough to add/ change data to update levels.

From the counterfeiter’s point of view: 1. brand popularity, 2. product popularity, 3. ease of
duplication, 4. ease of entry (due to active or passive brand protection from industry). // I would also
like to see the impact the economy has on the counterfeit business. I would imagine that business
(counterfeit) would be stronger than ever?

The model might serve well to consider other factors besides pure counterfeiting, e.g. tampering,
diversion, over labeling, up labeling, etc. All can play a role in feeding the counterfeit supply chain.
I am constantly challenged by no academic work in this counterfeit risk area - that is challenged by
managers asking for methods and for my trying to develop the tools. There are very generally
accepted attempts at the analysis from other disciplines, but no theoretical justification.

I like the approach as people often look beyond the qualitative aspects and focus on data
driven/quantitative information to make informed decisions.

We run a relative risk (with severity). I definitely like the idea of thinking things through holistically
across all industries - the retailers will all want one system.

It may be a logical approach but you are attempting to measure human motivation to commit a crime,
which is ofien discounts/ignores logical factors

I generally agree with mathematical models in predicting or quantifying an outcome, but I would like
to get more information on the rational for the determination of these various factors and sub-factors
and how this influences the outcome

See comment for question 3 (This shows where we would want to protect against and not necessarily
where we would be counterfeited)

Great place to start. . ..In all my dealings with a lot of brand managers, this item still seems to be low on
the priority list. It seems like consumer good companies that have high priced items are aware and try
to take measures (put security tags on products for example), but below a certain price point it falls
below package appeal for the consumer and product protection.

149

APPENDIX D — Case Study Roll-Up Detail

This appendix includes the Counterfeit Product Risk Model background

spreadsheets that were used in the Case Study. The data rolls-up from “III. Sub-F actor

Derivation” up to the “1. Overall Counterfeit Product Risk.”

Table 21. Case Study Overall Counterfeit Product Risk, Summary

Case Study: Confidential Food Company

 

 

 

 

1. Overall Counterfeit Risk

Bolded Box = Data Entry

 

Probability

 

Very High

"Eb

.—

:l:

Medium
Low

Very Low

 

Total Rank

 

 

 

 

 

 

(1) 95+% of product is manufactured at one
proprietary location, (2) The company has
comparably low volume combined with
advanced packaging and product features
that are somewhat complex to duplicate, (3)
the company monitors product in the

 

marketplace including online C2C

 

Factors

Summary Function: A+B+C+D+E=F

l

I All equal, manual derivation

 

l. Counterfeit-History

No known counterfeits, little diversion, but
related lower premium product are faked

 

2. Counterfeit-Ability

The product and from are frequently
counterfeited, but uniqre feature reduce risk

 

3. Counterfeit-
Attractiveness

This is a premium product with $2-5x
generic pricing, but tight supply chain
controls

 

4. Counterfeit-Hurdles

High hurdles for counterfeiters, in terms of
pkg features, tight supply chain, and
monitorinL

 

5. Market Profile

 

 

 

 

 

 

Brand loyal consumers, price insensitive,
pursue reputable sources, and in-house
production

 

 

150

 

 

Table 22. Case Study Factor 1 Counterfeit-History, Summary
11. Factors and Sub-Factors

l:l Bolded Box = Data Entry
Factor 1. Counterfeit-History

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'55 B

E f, .5 a 3

r E g 3 s

> >

No lmown counterfeits, little diversion, but related 1

Total Rank X lower premium product are faked
Summary Function: I I
A+B+C=D . All equal, manual entry
1.1 Known
Counterfeits 0 O O 0 X No Known Counterfeit Incidents
1.2 Known
Diversion! Pallets 0 O 0 0 X Some diversion, small scale, one-off
1.3 Related Product
Counterfeit O X 0 O 0 Related products are counterfeited

 

 

Table 23. Case Study Factor 2 Counterfeit-Ability, Summary

Factor 2. Counterfeit-Abili

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'5 3

E :5 .5 3 .3

s a ‘3 3 e

:> 2 >

The product and from are frequently counterfeited,
Total Rank X but unique feature reduce risk
[ Summary Function: A+B+C=D I 1 All equal, manual entry I

2.1 Product/ Bulk
Counterfeit X 0 0 0 0 Product and from counterfeited often
2.2 Package! Case 0 0 X 0 0 Other counterfeiting, but unique features
2.3 Supply/ Enter
Genuine Market 0 O 0 X 0 Difficult to entegenuine market en masse

 

 

 

 

 

 

 

 

151

 

Table 24. Case Study Factor 3 Counterfeit-Attractiveness, Summary
Factor 3. Counterfeit-Attractiveness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

in a, g 3

{I} .. 3

g a ‘3 3 g

> 2 :>

This is a premium product with $2-5x generic

Total Rank X pricinLMght supply chain controls
Summary Function: A+B+C=D j lAllgual, manual entry
3.1 Brand
Premium vs.
Fake Cost X 0 O 0 Premium product 52-5x generic
3.2 Supply
Chain/ B2B,
B2C, C2C 0 O 0 X 0 Very little in online marketplace, even C2C
3.3 Supply
Chain/
Monitoring 0 0 O X 0 Active monitoring and relationships

 

 

 

 

 

 

 

 

Table 25. Case Study Factor 4 Counterfeit-Hurdles, Summary
Factor 4. Counterfeit - Hurdles

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'5, 3

E 55, .5 3 .3

r ='= “g 3 s

> >

High hurdles for counterfeiters, in terms of pkg
Total Rank X features, tight supply chain, and monitoring
Summary Function: A +B+C=D J I All equal, manual entry

4.1 Package:
TE/CR
Features: 2+ 0 O 0 X 0 2 TFJCR on primary pkg, more on others
4.2 Package:
Printing/
Aesthetics 0 O 0 X 0 Complex, precision printing & special pkg
4.3
Marketplace
investigation 0 O O X 0 Product form counterfeited

 

 

 

 

 

 

 

 

152

 

 

 

Table 26. Case Study Factor 5 Market Profile, Summary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Factor 5. Market Profile
4"- B
.99 E
m g, .a 3 .3
r a ‘3 3 r
> 2 >
Brand loyal consumers, price insensitive,
pursue reputable sources, and in-house
Total Rank X production
1 Summary Function: A +B+C=D I I All equal, manual entry
5.1 Consume/
ggoing consume 0 0 O X 0 Brand loyalty, consumer pkg awareness
5.2 Consumer
Price Sensitivity 0 0 0 X 0 Brand loyalty, pursue reputable sources
5 .3 Contract
Manufacturing/
promo 0 0 0 X 0 95% production in-house, 100% promo/ kits

 

 

 

 

 

 

 

 

Table 27. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.1
III. Sub-F actor Derivation

A. Factor 1. Counterfeit-History

Factor 1. Counterfeit - History] Sub-Factor 1.1
[1.] Known Counterfeits

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

153

 

Total Rank
Very Very
HigL High Medium Low Low
No Known Counterfeit
O O 0 0 X Incidents
LSummary Function (for exam =B
Very High tbd
:5 nigh tbd
g Medium tbd
5 [DW tbd No known counterfeit
Very Low tbd X incidents

 

 

Table 28. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.2
Factor 1. Counterfeit - History/ Sub-Factor 1.2
L12 Known Diversion/ Pallets

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
High High Medium Low Low
0 0 O O X Some diversion, small scale, one-off
I Summary Function (for exam 1e) =B
Very High tbd
:3 it}: tbd
3 Medium tbd
5 LOW tbd Some diversion, but not en masse/
Very Low tbd X not pallet or shipping container loads

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 29. Case Study Factor Derivation, Counterfeit-History, Sub-factor 1.3
Factor 1. Counterfeit - History/ Sub-Factor 1.3

 

I 1.3 Related Product Counterfeit

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
_High High Medium Low Low
0 x o 0 0 Related products are counterfeitedJ
I Summary Function (for exam 1e) A=
Very High tbd
Q Standard quality product
3 _High tbd X
-° Medium tbd
5 Low tbd
Very Low tbd

 

 

 

 

 

 

 

Table 30. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.1

B. Factor 2. Counterfeit — Ability
Factor 2. Counterfeit - Ability/ Sub-Factor 2.1
[2.1 Product/ Bulk Counterfeit

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
H'gh High Medium Low Low
X 0 O O 0 Product and from counterfeited oftenJ
I Summagy Function (for exam Ie) =3
3‘ Very High tbd X
E flgh tbd Product and form counterfeited ofien,
_8 Medium tbd Two Product forms - 80% in a simple
2 LOW tbd but with very recognizable coloring
Very Low tbd and printing

 

 

 

 

154

 

 

 

 

 

Table 31. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.2
Factor 2. Counterfeit - Ability/ Sub-Factor 2.2

 

[2_.2 Package/ Case

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
High High Medium Low Low
0 O X 0 O Other counterfeiting, but unique featuresj
I Summary Function (for exam le) =B
3‘ Very High tbd
i5. High tbd
3 Medium tbd X Product and form counterfeiting common,
E 10‘” tbd but specific features lead to more
Very Low tbd challenge

 

 

 

 

 

 

 

Table 32. Case Study Factor Derivation, Counterfeit-Ability, Sub-factor 2.3
Factor 2. Counterfeit - Abllity/ Sub-Factor 2.3

 

I 2.3 Supply/ Enter Genuine Market

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
High High Medium Low Low
0 0 O X 0 Difficult to entegenuine market en masse I
I Summary Function (for exam le) =B
b VerLHigh tbd
a High tbd
.o .
.8 Medium tbd Single cases of product could be swapped
E LOW tbd X out by insiders, but most product is shipped
Very Low tbd direct to distributors or retailers

 

 

 

 

 

 

 

Table 33. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.1

C. Factor 3. Counterfeit - Attractiveness
Factor 3. Counterfeit - Attractiveness] Sub-Factor 3.1

 

I 3.1 Brand Premium vs. Fake Cost

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
iigh High Medium Low Low
X 0 O O 0 Premium product $2-5_xgeneric I
I Summary Function (for exam le) A=
Very @h tbd X
5 1m. tbd
g Medium tbd
E Low tbd
Very Low tbd This is a premium brand with $2-5x generic

 

 

 

 

 

 

155

 

 

 

 

Table 34. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.2
Factor 3. Counterfeit - Attractiveness/ Sub-Factor 3.2
I 3.2 Supply Chain/ 3213, B2C, C2C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
_Iiigh High Medium Low Low
0 o o x 0 Very little in online marketplace, even czcfl
I Summary Function (for exam le) A=
Very High tbd
5 _High tbd
3 Medium tbd
E LOW tbd X Very little product on general B2B, B2C,
Very Low tbd and C2C online markets

 

 

 

 

 

 

 

 

 

Table 35. Case Study Factor Derivation, Counterfeit-Attractiveness, Sub-factor 3.3
Factor 3. Counterfeit - Attractiveness! Sub-Factor 3.3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[3.3 Supply Chain/ Monitorig
Total Rank
Very Very
_High High Medium Low Low
0 0 O X 0 Active monitoring and relationships I
Summary Function (for exam le) A=
Very High tbd
5 Bligh tbd
3 Medium tbd
5 LOW tbd X Rigorous supply chain monitoring, quality
Very Low tbd control checks, distributor relationships etc.

 

 

 

 

 

 

 

 

 

Table 36. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.1
D. Factor 4. Counterfeit - Hurdles

Factor 4'. Counterfeit - Hurdles/ Sub-Factor 4.1

[4.1 Package: TE/CR Features: 2+

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
iigh @h Medium Low Low
0 O O X 0 2 TE/CR on primary pkhmore on others 3
I Summary Function (for exam le) =B
Very High tbd
5 are tbd
g Medium tbd
E LOW tbd X 2 TE/ CR features on primary package plus
Very Low tbd 2 more on other packaging

 

 

 

 

 

 

156

 

 

 

 

Table 37. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.2

Factor 4. Counterfeit - Hurdles/ Sub-Factor 4.2
{4.2 Pachage: Printing/ Aesthetics

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Rank
Very Very
High High Medium Low Low
Complex, precision printing &
O O O X 0 special pkg
Summary Function (for exam le) =B
3' Very High tbd
a High tbd
.o .
.8 Medium tbd Complex, precise, vibrant printing as
5 LOW tbd X well as a unique and very high
Very Low tbd quality bottle

 

 

 

 

 

 

Table 38. Case Study Factor Derivation, Counterfeit-Hurdles, Sub-factor 4.3

Factor 4. Counterfeit - Hurdles/ Sub-Factor 4.3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I 4.3 Marketplace investigation
Total Rank
Very Very
High High Medium Low Low
0 0 O X 0 Product form counterfeited
I Summary Function (for exam le) A=B
Very list; tbd
5 Hi h tbd
«1 __g . Two Product forms - 80% in a simple
3 Medrurn tbd but with very recognizable coloring
E LOW tbd X and printing, Active supply chain
Very Low tbd monitoring

 

 

 

 

 

 

Table 39. Case Study Factor Derivation, Market Profile, Sub-factor 5.1

E. Factor 5. Market Profile
Factor 5. Marketing/ Sub-Factor 5.1
FSJ Consume/ Ongoing consume

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

157

 

 

Total Rank
Very Very
High High Medium Low Low
Brand loyalty, consumer pkg
0 O 0 X 0 awareness
Summary Function =B
3‘ VerLHigh tbd
:3; flh tbd Long-term product, very brand loyal
.3 Medlum tbd consumers who are aware of the
E Low tbd X package/ printing (several examples
Very Low tbd of consumer contacts)

 

 

 

 

 

Table 40. Case Study Factor Derivation, Market Profile, Sub-factor 5.2
Factor 5. Marketing/ Sub-Factor 5.2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

£2 Consumer Price Sensitivity
Total Rank
Very Very
High High Medium Low Low
Brand loyalty, pursue reputable
O 0 O X 0 sources
Summary Function A=B
V H' tbd
Q Hiery 18h tbd the customers seek good deals but
‘1 —g , only from reputable sources. This is
.3 Medium tbd a high tech product and the high
,8 LOW tbd X percent of brand loyal customers are
Very Low tbd dedicated to the brand.

 

 

 

Table 41. Case Study Factor Derivation, Market Profile, Sub-factor 5.3
Factor 5. MarketiniSub—Factor 5.3
I 5.3 Contract Manufacturing promo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

158

 

 

Total Rank
Very Very
High High Medium Low Low
95% production in-house, 100%
0 0 0 X 0 promo/ kits
Summary Function =B
Very High tbd
:5 High tbd
g Medlum tbd 95% of production in-house and in one
5 low tbd X facility. 100% of promotional and
Very Low tbd special kits are assembled in-house

 

 

 

 

APPENDIX E — Case Study Example of
Automated Function

As was noted in the main text, considering the previous review of the nature of

the data and the warnings about keeping the analysis useable for the modelers, using a

manual (heuristic) process it supported, and preferable for few experiments. This

appendix includes an example of the final automated output (Table 42) and discussed the

summary of the total risk. Several definitions or clarifications are important:

The Total Rank is the overall probability of counterfeit risk for a specific product
that has been analyzed by an expert panel.

The probability is comprised of several categories, which include Very High,
High, Medium, Low, and Very Low.

The Factors are counterfeit attributes, such as “Counterfeit-History.”

The Rank Ranges are the numerical values that specify the categories, such as the
Size of the Rank Ranges and the Number of Rank Ranges. The Midpoints of
the Rank Ranges are also included.

The Summary Function formula used these variables, for example,
A+B+C+D+E=F.

For presentation simplicity, this experiment is conducted only on the final Total
Risk.

For presentation simplicity, the “X” was converted to a binary “l ,” and “0” (letter

O) was changed to a “0” (zero).

159

The automated summary of the total risk was determined by the following

process:

0 To start, the following background calculations were decided (Table 43):

0 Range: Size-Width

0 Range: Size-Number Range

0 Range: Midpoints

o The Summary Function formula used this formula and these variables:

A+B+C+D+E=F. Variable “A” represents Factor 1, variable “B” represents

Factor 2, and so on. Variable “F” is the probability rank for the Total Rank, which

was the average of the probability rank of the five factors.

Table 42. Case Study Example of Automated Total Risk Determination

 

 

 

 

 

Categories Very High High Medium Low (L) Very Low
(VH) M) (M) (VL)
Range: Size-Width (out of five) 0.25 1.5 1.5 1.5 0.25
Range: Size-Number Range 4.75-5.0 3.25-4.75 1.75-3.25 0.25-1.75 0-0.25
Range: Midpoint 4.875 4.0 2.5 1.0 .125

 

 

 

 

 

 

160

 

 

Table 43. Model Development — Example of Automated Summary of Total Rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Probability
Very . . Very
H' h High Medrum Low Low
Use a "Conditional
Formula" (essentially
Total Rank X an "if/ then" statement)
to place the "Total
Rank"
The summary for the
five factors, of the rank
. multiplied by the mid-
Summary F unctron (for .
= 1.3 pornt of the Rank
example) A+B+C+D+E F Range:
VH"‘4.875+H"4+M"2.5
+L*1+VL"'. l 25
1. Counterfeit- 1
History
2. Counterfeit- l
Abrlrty . Note: for presentation
3. Counterfert— l simplicity, the "X" was
Attractiveness converted to a "l "
4. Counterfeit- l
Hurdles
5. Market Profile 1
Background
Calculations
Probability
Very . . Very
High Hrgh Medrum Low Low
Determined by an
Expert Panel, this is
. selection is logical
if: 2: Rank 0.25 1.5 1.5 1.5 0.25 since Very High and
g Very Low are intended
to represent very close
to 9% and 100%.
Number of Rank 325- 1.75- 0.25-
Ranges 5'0'4'75 4.75 3.25 1.75 0'0‘25
Midpornt of Rank 4.875 4 2.5 1 0,125
Ranges

 

 

 

 

 

 

 

than the other factors

Summary Function with Factor 1 rated as 5x more important

 

5A+B+C+D+E=F

 

161

 

 

 

REFERENCES

162

18 USC Sec. 1963, Title 18, Part 1, Chapter 96. (2006). CHAPTER 96 - RACKETEER
INFLUENCED AND CORRUPT ORGANIZATIONS, Sec. 1963. Criminal
Penalties. Retrieved. from http://uscode.house.gov/uscode—
cgi/fastweb.exe?getdoc+uscview+tl 7 t20+93 2+O++%28real%20property,%20incl
uding%20things%20growing°/0200n,%20affixed%2Oto,%20and%20found%20in
%201and;%20%29%20%20%20%20%20%20%20%20%20%20.

21 CF R 120. (2003). Code of Federal Regulations, TITLE 21--FOOD AND DRUGS,
CHAPTER I--FOOD AND DRUG ADMINISTRATION, DEPARTMENT OF
HEALTH AND HUMAN, SERVICES, PART 120-—HAZARD ANALYSIS AND
CRITICAL CONTROL POINT (HACCP) SYSTEMS. Retrieved February 17,
2007, from http://www.cfsan.fda.gov/~lrd/FCF120.html

Acheson, David. (2007a). Food Protection, Food Safety and Food Defense. Paper
presented at the Association of Food and Drug Officials (AF DO) Annual

Conference. from http://www.afdo.ory afdo/Conferences/upload/ 07061 9-Food-
l 530%20-

%20Acheson%2ONEW%20AFDO%20Food%2OProtection%2OJune%202007_07
0615%20-%20Final.pdf.

Acheson, David. (2007b). Vulnerability Assessment Tools. Paper presented at the Global
Food Safety and Quality Conference. from
http://members.ifi.org/NR/rdonlyres/A3 748A89-71 58-4Al 7-BBFF-
26ACE9697ABC/O/GlobalFoodSafety_Acheson2.pdf.

Alcock, Lucy, Chen, Paul, Ch'ng, Hui Min, Hodson, Sarah, & et a1. (2003).
Counterfeiting: Tricks and trends. Journal of Brand Management, 11(2), 133.

Allan, John. (2006). A Heuristic Risk Assessment Technique for Birdstrike Management
at Airports. Risk Analysis, 26(3), 723.

Andel, Tom. (2006, July 1). Counterfeit Fighters. Paperboard Packaging, 91, 20-25.
Arrui‘iada, Benito, & Vazquez, Xosé H. (2006). When Your Contract Manufacturer

Becomes Your Competitor, Best Practice. Harvard Business Review, 84(19), 135-
144.

Aven, Terje. (2003). Foundations of Risk Analysis. Hoboken, NJ: John Wiley & Sons,
Inc.

Balfour, Frederik. (2005). Fakes! The global counterfeit business is out of control,

targeting everything from computer chips to life-saving medicines. It's so bad that
even China may need to crack down. Business Week(39l9), 54-64.

163

 

Banks, Erik. (2005). Catastrophic Risk: Analysis and Management. Hoboken, NJ: John
Wiley & Sons, Inc.

Bialik, Carl. (2007a). Efforts to Quantify Sales of Pirated Goods Lead to Fuzzy Figures
[Electronic Version]. Wall Street Journal On-line. Retrieved February 19, 2007
from <http://online.wsj.com/article/SB1 19274946863264] 17.html>.

Bialik, Carl. (2007b, October 19). THE NUMBERS GUY: Efforts to Quantify Sales of
Pirated Goods Lead to Fuzzy Figures. The Wall Street Journal, p. B1.

Bix, Laura, Clarke, Robb, Lockhart, Hugh, Twede, Diana, & Spink, John. (2007). Global
Data Standards in the Healthcare Supply Chain: The Business Case, prepared for
The G8] Healthcare Users Group (HUG). Michigan State University, School of
Packaging, May 30, 2007.

Bogdanich, Walt. (2007a, June 17). F.D.A. Tracked Poisoned Drugs, but Trail Went Cold
in China. New York Times.

Bogdanich, Walt. (2007b, December 17). Toxic Pipeline, Counterfeit Drugs’ Path Based
by Free Trade Zones. New York Times.

Bogdanich, Walt. (2008, March 20). Heparin Find May Point to Chinese Counterfeiting
New York Times.

Bogen, Kenneth T., & Jones, Edwin D. (2006). Risks of Mortality and Morbidity from
Worldwide Terrorism: 1968-2004. Risk Analysis, 26(1), 45.

Booth, Martin. (1999). The Dragon Syndicates: The Global Phenomenon of the T riads.
London: Bantam.

Brackett, Robert. (2004). Threat/ Vulnerability Assessment for Foods. Paper presented at
the 2004 FDA SCIENCE FORUM, FDA Science: The Critical Path fiom Concept
to Consumer, FDA Science Board Advisory Committee Meeting. fiom
http://www.fda. gov/ ohms/dockets/ ac/ 03/ slides/4001 s 1 _02_Brackett%20.ppt.

Branch, Ken. (2008). Brand Protection Strategy 2008. Paper presented at the Product
Authentication and Brand Security Conference (PABSO8). from Hardcopy.

Brekke, Brad. (2008, June 21). Organized Retail Theft: Fostering a Comprehensive
Public-Private Response. Loss Prevention, 11, 36.

Broder, James F. (2000). Risk Analysis and the Security Survey (2nd ed.). Boston:
Butterworth-Heineman.

Buchanan, Robbert, Dennis, S., & Miliotis, M. (2004). Initiating and Managing Risk

Assessments within a Risk Analysis Framework: FDA/CFSAN’S Practical
Approach. Journal of Food Protection, 67(9), 2058-2062.

164

 

Buchanan, Robert. (2004). Principles of risk analysis as applied to microbial food safety
concerns. Mitteilungen aus Lebensmitteluntersuchung und Hygiene, 95(1), 6-12.

Buchanan, Robert. (2007). Tools for Prioritizing Food Safety Concerns: An FDA
Perspective. Paper presented at the Tools for Prioritizing Food Safety Concerns
Workshop. from <http://www.fifsan.umd.edu/tools_2007.hhn>.

Buchanan, Robert. (2008). Equivalence and International Comparison of Food Safety
Systems - WHO Perspective, Presentation. Paper presented at the Annual
Conference, Symposium 81] Best Practices in Global Food Export and Import.
from Not Available.

Budima, Gjeneza. (2006). Can corruption and economic crime be controlled in
developing economies, and if so, is the cost worth it? Journal of Financial Crime,
13(4), 408.

Bullock, Karen, & Tilley, Nick. (2003). Crime reduction and problem-oriented policing:
Cullompton : Willan.

Burgan, Robert E. (2006). Fuel Models and Fire Potential from Satellite and Surface
Observations [Electronic Version]. USDA/US Forest Service, Rocky Mountain
Research Station,. Retrieved June 20, 2006 fi'om
http://www.fs.fed.us/land/wfas/firepot/fpipap.htm.

Bush, Laura. (2005). Q9 Establishes Common Terminology for Risk Management.
Pharmaceutical Technology, 29(9), 1.

Busta, F.F., & Chaisson, CF. (1997). Food Chemical Risk Analysis, Chapter 14 - The
philosophy of food chemical risk management. London: Blackie Academic &
Professional.

C-STAT, Center for Statistical Training and Consulting MSU. (2008a). Review of
Research Statistical Methods - Analysis of Results. In J. Spink (Ed.) (Spink
Presentation and C-STAT Recommendations ed.). East Lansing, Michigan: Tian
Song, MSU Center for Statistical Training and Consulting (C-STAT).

C-STAT, Center for Statistical Training and Consulting MSU. (2008b). Review of
Research Statistical Methods - Analysis of Surveys. In J. Spink (Ed.) (Spink
Presentation and C-STAT Recommendations ed.). East Lansing, Michigan: Paul
Curran, MSU Center for Statistical Training and Consulting (C-STAT).

C-STAT, Center for Statistical Training and Consulting MSU. (2008c). Review of
Research Statistical Methods - Set-up. In J. Spink (Ed.) (Spink Presentation and
C-STAT Recommendations ed.). East Lansing, Michigan: Paramita Chakraborty,
MSU Center for Statistical Training and Consulting (C-STAT).

Capra, Steven, & Canale, Raymond. (1998). Numerical Methods for Engineers (3rd ed.).
Boston.

165

CBER/FDA, Center for Biologics Evaluation and Research. (2006). Guidance for
Industry - Q9 Quality Risk Management [Electronic Version]. U. S. Food and
Drug Administration, Center for Drug Evaluation and Research (CDER), Center
for Biologics Evaluation and Research (CBER), , June 2006 fi'om
http://www.fdagov/Cber/gdlns/ichq9risk.htm.

CDRH/FDA, Center for Device and Radiological Health. (2009). Unique Device
Identification (UDI). Retrieved January 12, 2009, from
http://www.fda.gov/cdrh/ocd/udi/

Center for Problem-Oriented Policing. (2006, 6/20/06). What is POP? Retrieved May
24, 2007, fi'om http://www.popcenter.org/

Center for Strategic and International Studies. (2005). Security Controls on Scientific
Information and the Conduct of Scientific Research [Electronic Version]. A White
Paper of the Commission on Scientific Communication and National Security,
Homeland Security Program, Center for Strategic and International Studies
(CSIS). Retrieved June 2005 from
<http://www.csis.org/media/csis/pubs/0506_cscans.pdf>.

Cervone, H. Frank. (2006). Project risk management. OCLC Systems and Services, 22(4),
256.

CFSAN/F DA, Center for Food Safety and Applied Nutrition. (2001a). HACCP: A State-
of-the-Art Approach to Food Safety [Electronic Version]. Retrieved October from
http://www.cfsan.fda.gov/~lrd/bghaccp.html

CFSAN/F DA, Center for Food Safety and Applied Nutrition. (2001b, September 2001).
The A to Z, List of Terms. CFSAN/FDA Website Retrieved February 5, 2007,
from <http://www.cfsan.fda.gov/~dms/a2z-g.html>

CFSAN/F DA, Center for Food Safety and Applied Nutrition. (2005). GUIDANCE FOR
INDUSTRY, Submitting Requests, Under 21 CFR 170.39, Threshold of
Regulation For Substances Used in Food-Contact Articles [Electronic Version].
U. S. Food and Drug Administration, Center for Food Safety and Applied
Nutrition, Oflice of Food Additive Safety. Retrieved April from
http://www.cfsan.fda.gov/~dms/torguid.html.

CFSAN/F DA, Center for Food Safety and Applied Nutrition. (2007a). Carver+Shock
Software Tool, Home Page. Retrieved June 16, 2007, from
http://www.cfsan.fda.gov/~dms/carver.html

CFSAN/F DA, Center for Food Safety and Applied Nutrition. (2007b). An Overview of
the Carver Plus Shock Method for Food Sector Vulnerability Assessments
[Electronic Version]. U. S. Food and Drug Administration, Center for Food
Safety and Applied Nutrition from
<httD://www.fsis.usda.gov/PDF/CARVER.pdf>.

166

 

CFSAN/F DA, Center for Food Safety and Nutrition. (2002). Initiation and Conduct of
All 'Major' Risk Assessments within a Risk Analysis Framework [Electronic
Version]. U. S. Food and Drug Administration, Center for Food Safety and
Applied Nutrition. Retrieved 8/23/2007 from
http://www.cfsan.fda.gov/~dms/rafw-toc.html.

CFSAN/F DA, Center for Food Safety and Nutrition. (2003). Risk Assessment for Food
Terrorism and Other Food Safety Concerns [Electronic Version]. U. S. Food and
Drug Administration, Center for Food Safety and Applied Nutrition, Office of
Regulations and Policy from <http://www.cfsan.fda.gov/~dms/rabtact.html>.

CFSAN/FDA, Food and Drug Administration. (2006, July 17, 2006). ALERT: The
Basics. Food Defense and Terrorism, Center for Food Safety and Applied
Nutrition Retrieved February 5, 2008, from
<http://www.cfsan.fda.gov/~drns/alert.html>

CFSAN/F DA, Food and Drug Administration. (2007c). Guidance for Industry: Cosmetics
Processors and Transporters of Cosmetics Security Preventive Measures
Guidance [Electronic Version]. U.S. Food and Drug Administration, Center for
Food Safety and Applied Nutrition, November 2003 fi'om
<http://www.cfsan.fda. gov/~dms/secgui1 7.htm1>.

CF SAN/FDA, Food and Drug Administration. (2007d). Guidance for Industry: Dairy
Farms, Bulk Milk Transporters, Bulk Milk Transfer Stations and Fluid Milk
Processors: Food Security Preventive Measures Guidance [Electronic Version].
U.S. Food and Drug Administration, Center for Food Safety and Applied
Nutrition (CFSAN). Retrieved February 5, 2008 from
<http://www.cfsan.fda. gov/~drns/secgui l 6.html>.

CFSAN/F DA, Food and Drug Administration. (2007c). Guidance for Industry: Food
Producers, Processors, and Transporters: Food Security Preventive Measures
Guidance [Electronic Version]. U.S. Food and Drug Administration, Center for
Food Safety and Applied Nutrition. Retrieved February 5, 2008 fiom
<http://www.cfsan.fda.gov/~dms/secgui14.htrnl>.

Charlton, Stacy. (2004). CA CCP - Confinement Analysis and Critical Control Points
Approach to PMP Production. Paper presented at the USDA/APHIS Gene
Confinement Workshop. from
<http://www.aphis.usda.gov/brs/confine_workshop/charlton_present.pdf>.

Chaudhry, Peggy E. (2006). Changing levels of intellectual property rights protection for
global firms: A synopsis of recent U.S. and EU trade enforcement strategies.
Business Horizons, 49(6), 463-472.

Christ, Oliver P. (2005). Implementing ISO 14971:2000, Risk management standard has
global appeal. MEDICAL PRODUCT OUT SOURCING(J anuary/F ebruary), 46-48.

167

 

Clarke, Ronald V. (1997). Situational Crime Prevention, Chapter 1. Introduction. In R. V.

Clarke (Ed.), Situational Crime Prevention (2 ed.). Guilderland, NY: Harrow and
Heston Publishers.

Claycamp, H. G. (2006). Rapid Benefit-Risk Assessments: No Escape from Expert
Judgments in Risk Management. Risk Analysis, 26(1), 147-156.

Claycamp, H. G., & Hooberman, B. H. (2004). Antimicrobial Resistance Risk
Assessment in Food Safety. Journal of Food Protection, 67, 2063-2071 .

Claycamp, H.G. (2003). A contemporary approach to risk analysis in GFI #152
[Electronic Version]. FDA Veterinarian, 18, 2-5 from
http://www.fda.gov/cvm/index/fdavet/2003/jan-feb03.pdf

Clemens, Pat L., & Swallom, Donald W. (2005). Summing Risk - An International
Workshop and Its Results. Journal of System Safety, 41(6), 21.

Clinton, James M. (2008). REGULATIONS & STANDARDS, Determining the key
initial steps in risk management [Electronic Version]. In Vitro Diagnostics
Technology (I VDI), 16. Retrieved 11/5/2008 fi‘om
<http://www.devicelink.com/ivdt/archive/08/09/010.html>.

Closs, David J ., & McGarrell, Edmund F. (2004). Enhancing Security Throughout the
Supply Chain [Full Report]. East Lansing, MI: Michigan State University.

Coalition Against Counterfeit and Piracy. (2006). No Trade in Fakes, Brand Integrity
Tool Kit US Chamber of Commerce, Coalition Against Counterfeit and Piracy
(CACP).

Coffin, Bill. (2002). Forecasting Terrorism Losses. Risk Management, 49(11), 8.

Cohen, Jacob. (1960). A coefficient of agreement for nominal scales. Educational and
Psychological Measurement, 20(1), 37-46.

Collins, David. (2004). Brand Protection. Leatherhead, Surrey, UK: Pira International.

Congressional Research Service. (2003). Balancing Scientific Publication and National
Security Concerns: Issues for Congress [Electronic Version]. Congressional
Research Service (CRS), The Library of Congress, CRS Report for Congress,
Received through the CRS Web, Order Code RL31695. Retrieved July 9, 2003
from <http://www.fas.org/programs/ssp/bio/resource/documents/RL3 1695.pdf>.

Cook & Thurber. (2006). Quality, Safety and Security Expectations and Criteria for
Manufacturing Facilities of Packaging Materials and Nonfood Items (Revised
February 2006 ed.). Ann Arbor, Michigan: NSF International.

Council of Europe. (2006). Counterfeit Medicines: Survey Report. Strasbourg: Council of
Europe Publications.

168

 

Counterfeiting Intelligence Bureau. (1997). Countering Counterfeiting. A guide to
protecting & enforcing intellectual property rights: Counterfeiting Intelligence
Bureau (CIB), International Chamber of Commerce (ICC).

Counterfeiting Intelligence Bureau. (2007). Overview of Counterfeiting. Retrieved April
11, 2007, fi'om http://www.icc-ccs.org/cib/overview.php

Cramer, Duncan, & Howitt, Dennis. (2004). The Sage Dictionary of Statistics. Thousand
Oaks, CA: Sage Publications.

Cruz, Marcelo G. (2002). Modeling, Measuring and Hedging Operational Risk. New
York, NY: John Wiley & Sons, Inc.

Dain, Steven. (2002). Normal Accidents: Human Error and Medical Equipment Design.
The Heart Surgery Forum, 5(3), 254-257.

deKieffer, Donald. (2006a). Potential Liability for Counterfeit Medications. Journal of
Pharmacy Practice, 19(4).

deKieffer, Donald. (2006b). Trojan drugs: counterfeit and mislabeled pharmaceuticals in
the legitimate market. American Journal of Law and Medicine, 32(2-3), 325-349.

Department of Homeland Security. (2007). Agriculture and Food Critical Infiastructure
and Key Resources Sector-Specific Plan, as input to the National Infrastructure
Protection Plan.

DHS, Department of Homeland Security. (2007). Strategy to Enhance International
Supply Chain Security [Electronic Version]. Department of Homeland Security,
DHS/ HSDL fi'om http://www.dhs.gov/xlibrary/assets/plcy—
internationalsupplychainsecuritystrategypdf

Dietrich, Ed, Puskar, Erik, Grace, Andy, Allen, Mary Ann, & Schmitt, George. (2006).
Considering RFID for Use in the Fight Against Counterfeiting: Coalition Against
Counterfeiting and Piracy (CACP), U.S. Chamber of Commerce.

DiLonardo, Robert L., & Clarke, Ronald V. (1996). Reducing The Rewards Of
Shoplifting: An Evaluation Of Ink Tags. Security Journal, 7(1), 11.

Duffie, Darrell, & Pan, Jun. (1997). An Overview of Value at Risk. Journal of
Derivatives, 4(3).

Eban, Katherine. (2005). Dangerous Doses: How Counterfeiters are Contaminating
America's Drug Supply (lst ed.). NY, NY: Harcourt, Inc.

Eisenreich, Giinther, & Sube, Ralf. (1982). Dictionary of mathematics (V 01. 1). New
York: Elsevier North Holland, Elsevier Scientific Pub. Co.

169

Elsworth, J eff. (2007). Guest Lecture, VM821 Food Protection and Defense, Michigan
State University, National Food Safety & Toxicology Center, 9/15/2007.

Ericson, Richard V. (2006). Ten Uncertainties of Risk-Management Approaches to
Security. Canadian Journal of Criminology and Criminal Justice, 48(3), 345.

Etherton, John. (2007). Industrial Machine Systems Risk Assessment: A Critical Review
of Concepts and Methods. Risk Analysis, 27(1), 71-82.

Etherton, John, Main, Bruce, Cloutier, Dennis, & Christensen, Wayne. (2008). Reducing
Risk on Machinery: A Field Evaluation Pilot Study of Risk Assessment. Risk
Analysis, 28(3), 711-721.

European Commission. (2005). Customs: Commission and Member States step up action
against counterfeiting and piracy, PRESS RELEASE [Electronic Version].
European Commission (EC), Reference: IP/05/14 7, Brussels. Retrieved February
17, 2007 fi'om
<http://europa. eu/rapid/pressReleasesAction.do?reference=IP/0 5/ 147&format=H
TML&aged=0&language=en&guiLanguage=en>.

European Commission. (2006). Summary of Communicy Customs Activities on
Counterfeit and Piracy, Results at the European Border - 2006 [Electronic
Version]. European Commission - Taxation and Customs Union (EC/TCU) fi'om
http://ec.europa.eu/taxation_customs/resources/documents/customs/customs_cont
rols/counterfeit_piracy/statistics/counterf_comm_2006_en.pdf.

European Commission. (2007). Counterfeit and Piracy [Electronic Version]. European
Commission - Taxation and Customs Union (EC/TCU) fiom
http://ec.europa.eu/taxation_customs/customs/customs_controls/counterfeit_pirac
y/combating/index_en.htm.

European Union. (2003). Council Regulation (EC) No 1383/2003. Ofi‘icial Journal of the
European Union, 46(L 196), 7-11.

FDA, Food and Drug Administration. (1999). Workshop on International Outreach and
Training on Good Agricultural and Good Manufacturing Practices for Fresh
Produce, Notice of Meeting. Retrieved. fiom
<http://www.fda.gov/ohrrns/dockets/98fi'/032599d.txt>.

FDA, Food and Drug Administration. (2002). A Safe Food Supply [Electronic Version].
Performance Plan and Summary U. S. Food and Drug Administration from
http://www.fda.gov/ope/fy02plan/food.html.

FDA, Food and Drug Administration. (2003). Counterfeit Drug Task Force Interim
Report [Electronic Version] from
http://www.fda.gov/oc/initiatives/counterfeit/report/interim_report.html.

170

 

FDA, Food and Drug Administration. (2004). Combating Counterfeit Drugs [Electronic
Version]. U.S. Food and Drug Administration. Retrieved February 2004 from
<http://www.fdaflv/oc/initiatives/counterfeit/report02_04.html>.

FDA, Food and Drug Administration. (2006a). FDA Counterfeit Drug Task Force
Report: 2006 Update [Electronic Version]. Retrieved June 8, 2006 from
http://www.fda.gov/oc/initiatives/counterfeit/report6_06.htrrrl.

FDA, Food and Drug Administration. (2006b). Prescription Drug Marketing Act
Pedigree Requirements; Effective Date and Compliance Policy Guide; Request
for Comment. Retrieved. from
<http://www.fda.gov/OHRMS/DOCKETS/98fr/06-5362.htm>.

FDA, Food and Drug Administration. (2007a). Counterfeit Drugs Questions and
Answers. Retrieved February 5, 2008, from
http://www.fda.gov/oc/initiatives/counterfeit/qa.html

FDA, Food and Drug Administration. (2007b). FDA Publishes Counterfeit Toothpaste,
Firm Press Release [Electronic Version]. Retrieved February 5, 2008 from
<http://www.fda.gov/oc/m/firmrecalls/colgate06 07.html>.

Federal Bureau of Investigation. (2006). Criminal Investigation Handbook for
Agroterrorism.

Felson, Marcus. (1998). Crime and everyday life / Marcus F elson. (2nd ed. ed.):
Thousand Oaks, Calif. : Pine Forge Press, c1998.

Felson, Marcus, & Clarke, Ronald V. (1997). Business and Crime Prevention. New York,
NY: Criminal Justice Press. .

Feys, Robert, & Fitch, Frederic B. (1969). Dictionary of symbols of mathematical logic
(Vol. xiii). Amsterdam: North-Holland Pub. Co.

Fitzmaurice, G. (2002a). Measurement error and reliability. Nutrition, 18(1), 112-114.

Fitzmaurice, G. (2002b). Statistical methods for assessing agreement. Nutrition, 18(7-8),
694-696.

Fleiss, J .L. (1971). Measuring nominal scale agreement among many raters.
Psychological Bulletin, 76(5), 378-382.

Fleiss, J.L., & Cohen, J. (1973). The equivalence of weighted kappa and the intraclass
correlation coefficient as measures of reliability. Educational and Psychological
Measurement, 33(2), 613-619.

Food and Agriculture Organization of the United Nations. (2002). Good Agricultural
Practices (GAPS) [Electronic Version], June. Retrieved March 21, 2008 fiom
http://www.fao.org/wssd/sard/documents/faogapen.doc.

171

Fotis, Nick, & Bix, Laura. (2006). Sample Size Selection Using a Margin of Error
Approach [Electronic Version]. Medical Device & Diagnostic Industry (MDDI),
COVER STORY, 80. Retrieved 11/5/2008 from
<http://www.devicelinl_(.com/mddi/archive/06/1 0/002.html>.

Fourtou, J can-Rene. (2006, November 26). Opening Speech. Paper presented at the
BASCAP (Business Action to Stop Counterfeiting and Piracy) meeting, Paris.

F SIS/U SDA, Food Safety and Inspection Service. (1998). Salmonella Enteritidis Risk
Assessment [Electronic Version]. U. S. Department of Agriculture (USDA), Food
Safety and Inspection Service (FSIS) from <http://www.fsis.usda.gov/ophs/risk/>.

FSIS/USDA, Food Safety and Inspection Service. (2003a). Listeria Monocytogenes
Resources and Risk Assessment [Electronic Version] fi'om
<http://www.fsis.usda.gov/oa/topics/lm.htm>

FSIS/USDA, Food Safety and Inspection Service. (2003b). Risk Analysis, FSIS/ USDA,
Backgrounders/ Key Facts [Electronic Version], June 16, 2007. Retrieved July
2003, from <http://www.fsis.usda.gov/oa/backgr_ound/riskanal.htm>.

FSIS/USDA, Food Safety and Inspection Service. (2005, March 2005 ). BSE Resources
and Risk Assessment U. S. Department of Agriculture (USDA), Food Safety and
Inspection Service (F SIS), fi'om
http://www.fsis.usda.gov/Fact_Sheets/Bovine_Spongiform_Encephalopathy_Mad
_Cow_Disease/index.asp

Gerardi, Kristopher, Lehnert, Andreas, Sherlund, Shane, & Willen, Paul. (2008). Making
Sense of the Subprime Crisis [Electronic Version]. Brookings Papers on
Economic Activity. Retrieved September 11, 2008 fi'om
http://www.brookings.edu/economics/bpea/~/media/Files/Programs/ES/BPEA/20
08_fall_bpea_papers/2008_fall_bpea_gerardi_sherlund_lehnert_willen.pdf.

Global Business Leaders Alliance Against Counterfeiting. (2003). Product Counterfeiting
Overview. Global Counterfeiting, Background Document, Global Business
Leaders Alliance Against Counterfeiting (GBLAA C).

GlobalOptions Incorporated. (2003). An Analysis of Terrorist Threats to the American
Medicine Supply. Gaithersburg, Maryland: Signature Book Printing,
GlobalOptions Inc.

Goldstein, Herman (1979). Improving Policing: A Problem-Oriented Approach. Crime &
Delinquency, 25(2), 236-258.

Graver, PA. (2001). Process Hazard Analysis - Failure Mode Effects Analysis (FMEA)
[Electronic Version]. FMEA Information Center. Retrieved Rev. 1, 01/15/01 from
http://www.frneainfocentre.com/examples/shortintroexample.pdf.

172

 

Green, Robert T., & Smith, Tasman. (2002). Executive Insights: Countering Brand
Counterfeiters. Journal of International Marketing, 10(4), 89-106

Greenberg, Eric F. (2003, October 2003). Packaging, drug safety, terror and Counterfeits.
Packaging Digest, 40, 46(41).

Grocery Manufacturer's Association. (2007). Part I: Counterfeit consumer goods. GMA
Smartbrief Special Report.

Groupement International de la Repartition Pharrnaceutique. (2006). Advancing the
Single Market for Medicines. (GIRP)(European Association of Pharmaceutical
Full-Line Wholesalers).

Gwet, K. (2001). Handbook of inter-rater reliability: how to estimate the level of
agreement between two or multiple raters: Gaithersburg, MD: Stataxis.

Hallerbach, Winfried G., & Menkveld, Albert J. (2004). Analysing Perceived Downside
Risk: the Component Value-at-Risk Framework. European Financial
Management, 10(4).

Hammerbeck, Pauline. (2008). Protect and Deter. Brand Packaging Magazine, 12(5),
6(5).

Hassenzahl, David M. (2006). Implications of Excessive Precision for Risk Comparisons:
Lessons from the Past Four Decades. Risk Analysis, 26(1), 265-276.

Hazewinkel, M. (1988). Encyclopaedia of Mathematics: an Updated and Annotated
Translation of the Soviet "Mathematical Encyclopaedia ". Boston Dordrecht.

Hopkins, David M., Kontnik, Lewis T., & Tumage, Mark T. (2003). Counterfeiting
Exposed: How to Protect Your Brand and Market Share Hoboken, New Jersey:
John Wiley & Sons.

Howick, Susan, Eden, Colin, Ackermann, Fran, & Williams, Terry. (2008). Building
confidence in models for multiple audiences: The modelling cascade. European
Journal of Operational Research, 186(3), 1068-1083.

Imitating property is theft - Counterfeiting. (2003). The Economist, 36 7(8324), 70.

International Conference on Harmonization. (2005). Quality Risk Management - Q9.
Retrieved. from http://www.pmda.go.jp/ich/q/q9_06_9_1e.pdf.

INTERPOL. (2007). Intellectual Property Crimes: Trademark Counterfeiting &
Copyright Piracy. Interpol Intellectual Property Crimes Action Group (IIPCA G).

ISO, International Organization for Standardization. (2002). New ISO and IEC guide to
risk management terminology Press Release(Ref.: 833 ).

173

ISO, International Organization for Standardization. (2007a). Committee Draft of ISO
31000 “Risk management - Guidelines on principles and implementation of risk
management”

ISO, International Organization for Standardization. (2007b). Committee Draft of
ISO/IEC Guide 73 “Risk management - Vocabulary”.

ISO, International Organization for Standardization. (2008a). ISO/DIS 31000, Risk
management -- Principles and guidelines on implementation.

ISO, International Organization for Standardization. (2008b). ISO/DIS/IEC 31010 Ed.
1.0: Committee Review Document, Risk Management - Risk Assessment
Techniques.

Ivanova, Kate. (2007). Corruption, illegal trade and compliance with the Montreal
Protocol. Environmental and Resource Economics, 38(4), 475-496.

Jablonowski, Mark. (1994). Words or numbers? Risk Management, 41 (12), 47.

J ablonowski, Mark. (1996). Using probabilistic risk analysis to improve risk
management. Risk Management, 43(3), 23.

J ablonowski, Mark. (2005). Do Catastrophe Models Mislead? Risk Management, 52(7),
32.

Jain, Vineet Kumar, & Davidson, Rachel Ann. (2007). Application of a Regional
Hurricane Wind Risk Forecasting Model for Wood-Frame Houses. Risk Analysis,
27(1), 45-58.

J ohanning, Henrik, & Dahlgaard, Asger. (2007). Does quality risk management have
value? Pharmaceutical Technology Europe, 19(12), 12.

J otcharn, Richard B. (2005). Authentication, Antitarnper, and Track-and-Trace
Technology Options To Protect Foods. Journal of Food Protection, International
Association for Food Protection, 68(6), 1314-1217(13 14).

Kallman, James. (2007). Identifying Risk. Risk Management Magazine, 54(9), 58(52).
Kaplan, Stan. (1997). The Words of Risk Analysis. Risk Analysis, 1 7(4), 407-417.

Kara-Zaitri, C., Keller, A.Z., Barody, I., & Fleming, P.V. (1991). An improved FMEA
methodology. Paper presented at the Reliability and Maintainability Symposium.
fiom <http://ieeexplore.ieee.org/xpl/RecentCon.jsp?punumber=524>.

Karush, William, & Tarcov, Oscar. (1989). The Crescent Dictionary of Mathematics.
New York: Macmillan.

174

 

Kelly, Robert. (2005). Illicit T rafi‘icking: A Reference Handbook. Santa Barbara,
California: ABC-CLIO.

Kemshall, Hazel. . (2003). Understanding Risk in Criminal Justice: Buckingham : Open
University Press.

Kennedy, Jim. (2005). Defining a Successful Global Brand Protection Strategy. PIRA,
Innovations in Security Technology Conference.

Kiviat, Barbara. (2008). Reassessing Risk. Time, I 72(20), GB.1.

Kmenta, Steven, & Ishii, Kosuke. (2000). Scenario-Based FMEA: A Life Cycle Cost
Perspective Paper presented at the 2000 ASME Design Engineering Technical
Conferences. fi'om <http://www.medicalhealthcarefinea.com/papers/kmenta.pdf>.

Kontnik, Lewis. (2004). Pharmaceutical Counterfeiting: Preventing the Perfect Crime
[Electronic Version]. FFF Enterprises Incorporated, 2008 fi'om
<http://www.flfenterprises.com/F F F/Downloads/fff_wht _ppr_1 1 1204.pdf>.

Kontnik, Lewis T. (2006). Manufacturer and Industry Responses to the Counterfeiting
Challenge. Journal of Pharmacy Practice, 19(3), 140-145.

Kosal, Margaret E. (2006). Terrorism Targeting Industrial Chemical Facilities: Strategic
Motivations and the Implications for U.S. Security. Studies in Conflict and
Terrorism, 29(7), 719.

Krofi, Steve. (2004). Boosting For Billions [Electronic Version]. 60 Minutes, 2004.
Retrieved March 1, 2007 from
<httr)://www.cbsnews.com/stories/2004/02/20/60minutes/main601 396.§_html>.

Lam, James. (2003). Enterprise Risk Management: From Incentives to Controls. New
York, NY: John Wiley & Sons, Inc.

Langford, Ian H. (2002). An Existential Approach to Risk Perception. Risk Analysis,
22(1), 101-120.

Leahy, Patrick. (2008). Statement Of Senator Patrick Leahy (D-Vt.), Chairman, Senate
Judiciary Committee, On The Enforcement Of Intellectual Property Rights Act Of
2008. United States Sentate.

Lee, Charles R., & Shimpi, Prakash. (2005). The Chief Risk Officer: What Does It Look
Like and How Do You Get There? Risk Management, 52(9), 34.

Liang, B. A. (2006). Fade to black: Importation and counterfeit drugs. American Journal
of Law & Medicine, 32(2-3), 279-323.

Lipscomb, Brad, & Lewis, Austin. (2004). The Principles of Six Sigma: Building a
Quality Claims Management Program. Risk Management, 51(2), 30.

175

Los Angeles County Development Corporation. (2007). A False Bargain: The Los
Angeles County Economic Consequences of Counterfeit Products [Final Draft].
Los Angeles, CA.

Lukasik, Dave. (2008). Motorola, Motorola Inc. Brand Protection. Paper presented at the
Product Authentication and Brand Protection (PABSO8). from Hardcopy.

Martin, Andy. (2007). Industry consults on new regulations. Pharmaceutical Technology
Europe, 19(1), 18.

Marvel, Janet. (2008). Throw the Bums Out! Tips on Fighting Counterfeiting in Court
and Online. Paper presented at the Product Authentication and Brand Security
(PABSOS). from Hardcopy.

Marvin, H. J. P., Kleter, G. A., Frewer, L. 1., Cope, S., Wentholt, M. T. A., & Rowe, G.
(2009). A working procedure for identifying emerging food safety issues at an
early stage: Implications for European and international risk management
practices. Food Control, 20(4), 345-356.

 

McNeely, Susan. (2007). Selecting Brand-Security Technology Suppliers. Packaging
Digest, 44(3), 39.

McNeely, Susan. (2008). Protecting Health Care Products - Lessons Learned with Eli
Lilly. Paper presented at the Product Authentication and Brand Security
Conference (PABSOS). fi'om Hardcopy.

Merriam-Webster. (2004). The Merriam-Webster Dictionary -- New Edition.

Mil-Std-882D. (2000). MIL-STD-882D, 10 February 2000, SUPERSEDING MIL-STD-
882C, 19 January 1993, DEPARTMENT OF DEFENSE, STANDARD
PRACTICE FOR SYSTEM SAFETY, AMSC N/A AREA SAFT,

<hgp://safmcenter.nafl.mil/instructions/osh/milstd882d.flf>.

Millar, James A., & Yeager, Fred C. (2007). Issues in fighting financial crime. Economic
Aflairs, 2 7(1), 39.

Mitchell, Mark L. (1989). The Impact of External Parties on Brand-Name Capital: The
1982 Tylenol Poisonings and Subsequent Cases. Economic Inquiry, 27(4), 601.

Moorman, Marc. (2007). Industry Perspective on Food Safety (Zurich Hazard Analysis).
Paper presented at the NFSTC, VM810 Seminar.

Morris, Julian, & Stevens, Philip. (2006). Counterfeit medicines in less developed
countries: Problems and solutions. London: International Policy Network.

Murphy, Niall. (2006). REGULATORY OUTLOOK, Understanding Usability Standards
for Medical Devices [Electronic Version]. Medical Device & Diagnostic Industry

176

(MDDI), 74. Retrieved 11/5/2008 from
<http://www.devicelink.com/mddi/archive/06/05/01 1.html>.

Nader, Jihad S. (2002). The Manager's Concise Guide to Risk. Hoboken, NJ: John Wiley
& Sons, Inc.

Naim, Moises. (2005). Illicit: How Smugglers, T raflickers, and Copycats are Hijacking
the Global Economy. NY, NY: Doubleday.

Nakayachi, Kazuya. (1998). How Do People Evaluate Risk Reduction When They Are
Told Zero Risk Is Impossible? Risk Analysis, 18(3), 235-242.

National Research Council. (1996). Understanding Risk: Informing Decisions in a
Democratic Society. National Research Council (NRC), Washington, DC:
National Academy Press.

National Retail Federation. (2006). NRF: ORGANIZED RETAIL CRIME RAMPANT.
Women's Wear Daily (WWD), National Retail Federation (NRF), 191(123), 17.

New York City Comptroller's Office. (2004). Bootleg Billions: The Impact of the
Counterfeit Goods Trade on New York City, NYC Comptroller's Office,
November 2004.

O'Kane, Gerry, & Gopalkrishnan, Shanker. (2005). The Future of Anti-Counterfeiting,
Brand Protection and Security Packaging III. Leatherhead, Surrey, UK: Pira
International Ltd. (PIRA).

Occupational Safety & Health Administration. (1995, 07/26/1995 ). OSHA's approval of
the Zurich Hazard Analysis, as an appropriate method for conducting the process
hazard analysis which is required by 1910.119, Standard Interpretations. U.S.
Department of Labor, Occupational Safety & Health Administration (OSHA)
Retrieved June 16 2006, from

http://osha.gov/pls/oshaweb/owadisp.show_document?p_table=INTERPRETATI
ONS&p_id=21 867

OECD, Organization for Economic Co-Operation and Development. (2007a). The

Economic Impact of Counterfeiting and Piracy, Executive Summary [Draft
September 2007] .

OECD, Organization for Economic Co-Operation and Development. (2007b). The

Economic Impact of Counterfeiting and Piracy, Executive Summary, [Draft June
2007] .

OECD, Organization for Economic Co-Operation and Development. (2007c). The
Economic Impact of Counterfeiting and Piracy, Part I: Overall Assessment

[Draft].

177

 

OECD, Organization for Economic Co-Operation and Development. (2007d). The
Economic Impact of Counterfeiting and Piracy, Part II: Efforts to Combat
Counterfeiting and Piracy [DRAFT September 2007] [Electronic Version].
Organization for Economic Co-Operation and Development (OECD). Retrieved
November 22, 2007.

OECD, Organization for Economic Co-Operation and Development. (2007c). The
Economic Impact of Counterfeiting and Piracy, Part 111: Industry Sectors
[DRAFT September 2007] .

Office of Management and Budget. (2006). Proposed Risk Assessment Bulletin. fi'om
http://www.whitehouse.gov/omb/inforeg/proposed_risk_assessment_bulletin_010
906.pdf.

Onodera, K. (1997). Eflective techniques of FMEA at each life—cycle stage. Paper
presented at the Reliability and Maintainability Symposium. from
http://ieeexplore.ieee.org/xpl/RecentCon.jsp?punumber=4288>.

Paradise, Paul R. (1999). Trademark counterfeiting, product piracy, and the billion dollar
threat to the U.S. economy. Westport, Conn.: Quorum Books.

Parloff, R. (2006, May 1). Not exactly counterfeit. Fortune, 153, 108.

Patterson, Jacqueline, Meek, M. E., Strawson, Joan E., & Liteplo, Robert G. (2007).
Engaging Expert Peers in the Development of Risk Assessments. Risk Analysis,
27(6), 1609-1621.

Phillips, Anthony. (2008). Emerging Risk Analysis. Paper presented at the ASIS -
Lansing Chapter.

Phillips, Tim. (2005). Knockofl? the deadly trade in counterfeit goods : the true story of
the world ’s fastest growing crime wave. London: Kogan Page.

Pierce, Lisa McTigue. (2004). Building a solid GlaxoSmithKline: multiple mergers have
made GSK a master at controlling change, from its operations to country-specific
labeling. Food & Drug Packaging, 68(12), 22(28).

Pittinger, Charles A., Brennan, Thomas H., Badger, Drew A., Hakkinen, P. J., &
Fehrenbacher, M. Catherine. (2003). Aligning Chemical Assessment Tools
Across the Hazard-Risk Continuum. Risk Analysis, 23(3), 529-535.

Popping, Bert. (2002). The application of biotechnological methods in authenticity
testing. Journal of Biotechnology, 98(1), 107-112.

Prahl, Dennis S. (2007). Exhausted and Gray, but Still Going Strong: A Comparative
View of Parallel Imports fi'om the Trademark Perspective, 11. Preventing parallel
imports under trademark law. Retrieved October 12, 2007, fi'om
<http://www.la_(gs.wmflPPropMy/GmyMagkeUGrayMaOZ.html>

178

PriceWaterhouseCooper. (2007). Economic Crime: People, Culture and Controls
[Electronic Version]. The 4th biennial Global Economic Crime Survey
Investigations and Forensic Services, Price WaterhouseCooper (PWC) from
http://www.pwc.conr/extweb/pwcpublications.nsf/docid/1 E0890149345149E8525
73 7000705AF 1 /$fi1e/PwC_2007GECS.pdf.

Quaranta, Anna Grazia, & Zaffaroni, Alberto. (2008). Robust optimization of conditional
value at risk and portfolio selection. Journal of Banking & Finance, 32(10).

Rudolf, Paul M., & Bernstein, Ilisa B.G. (2004). Counterfeit Drugs. New England
Journal of Medicine, Volume 350:1384-1386, Number 14.

Rudolph, Harvey. (2003). REGULATORY OUTLOOK, Do We Need Medical Device
Risk Management Certification? [Electronic Version]. Medical Device &
Diagnostic Industry (MDDI). Retrieved 11/5/2008 fi'om
<http://www.devicelink.com/mddi/archive/03/1 1/001 .html.

Ryan, Richard. (2005). Risk Assessment to Drive Research for Contaminant Detection.
Paper presented at the First Annual Food Protection and Defense Conference,
Session: Food Processing, Decontamination, and Disposal. fiom
https://www.ift.org/fooddefense/6-Ryan.pdf

Saita, Francesco. (2007). Value at Risk and Bank Capital Management. New York:
Elsevier.

Savada, Andrea M., & Metz, Helen C. (1999). The Sociology and Psychology of
Terrorism: Who Becomes a Terrorist and Why? Federal Research Division,A
Report Prepared under an Interagency Agreement by the s September 1999
Author: Rex A. Hudson Editor: Marilyn Majeska Project Managers: Federal
Research Division (FRD) Library of Congress Washington, D. C. 20540-4840.

Scandia National Laboratory. (2007). A Safer Food Supply: Sandia and FDA to Make it
So [Electronic Version]. NEWS RELEASES, FOR IMMEDIA TE RELEASE from
<http://www.sandia.gov/news/resources/releases/2007/fda-carver.html>.

Schmidt, Mike W. (2004). Regulatory Outlook, The Use and Misuse of FMEA in Risk
Analysis [Electronic Version]. Medical Device & Diagnostic Industry (MDDI).
Retrieved 11/5/2008 fi'om
<http://www.devicelink.com/mddi/archive/O4/03/001 .html>.

Schwarz, Brian, & Wong, Vanessa. (2006). Counterfeit CURES [Electronic Version].
INSIGHT, Center for Medicine in the Public Interest (CMPI). Retrieved
OCTOBER 2006 from
http://cmpi.org/PDFs/conference_transcripts_reports/Counterfeit_Cures_A_Repor
t_From_the_American_Chamber_of_Commerce_in_Shanghai.pdf.

Shepherd, Richard, Barker, Gary, French, Simon, Hart, Andy, Maule, John, & Cassidy,
Angela (2006). Managing Food Chain Risks: Integrating Technical and

179

Stakeholder Perspectives on Uncertainty. Journal of Agricultural Economics,
57(2), 313-327.

Sim, J ., & Wright, C. C. (2005). The Kappa Statistic in Reliability Studies: Use,
Interpretation, and Sample Size Requirements. Physical Therapy, 85, 257-268.

Smith, Barbara-Helene. (2004). Regulatory Outlook, Developing a Comprehensive
Quality System [Electronic Version]. Medical Device & Diagnostic Industry
(MDDI). Retrieved 11/5/2008 fi'om
<http://www.devicelink.com/mddi/archive/04/06/021 .html>.

Smith, Michelle A. (1999). FDA ’s Good Agricultural Practices. Paper presented at the
Apple Cider Food Safety Workshop. from
http://www.cfsan.fda.gov/~comm/cidwsmi2.htrnl.

Society of Automotive Engineers (SAE). (2002). Potential Failure Mode and Effects
Analysis in Design (Design FMEA) and Potential Failure Mode and Effects
Analysis in Manufacturing and assembly Processes (Process F MEA). Reference
Manual, Document Number: SAE J I 739.

Sodipo, Bankole. (1997). Piracy and counterfeiting : GA TT, TRIPS and developing
countries. London: Kluwer Law International.

Spink, John. (2008). Meeting Notes - Imported Foods Vulnerability Project Kick-Ofl
Conference Call. East Lansing, Michigan: Michigan State University.

Spitzer, Robert L, & Fleiss, Joseph L. (1974). A Re-analysis of the Reliability of
Psychiatric Diagnosis. The British Journal of Psychiatry, 125, 341-347.

Stamatelatos, Michael. (2000). Probabilistic Risk Assessment: What Is It And Why Is It
Worth Performing It? Safety & Mission Assurance News - Mission Success Starts
With Safety, NASA(April 2000).

Standards New Zeeland. (2008). Release of Draft International Standard (DIS) for public
comment: ISO 31000 DIS - First draft international Standard on Risk
Management.

Studler, Nicholas. (2008). Brand and IP Protection at The Coca-Cola Company. Paper
presented at the Product Authentication and Brand Strategy (PABSO8). fi'om
Hardcopy.

Supreme Court of the State of Florida. (2003). IN THE SUPREME COURT OF THE
STATE OF FLORIDA, Case No: SC02-2645, FIRST INTERIM REPORT, OF
THE SEVENTEENTH STATEWIDE GRAND JURY, Case No: SC02-2645,
First Interim Report of the Seventeenth Statewide Grand Jury, Dec. 19, 2003.

Taylor, Charles. (2006). The RMA operational risk management framework. The RMA
Journal (Risk Management Association), 4(4).

180

 

Tibiletti, Luisa. (2008). Value-at-risk: is lacking in sub-additivity just an annoying
technicality? International Journal of Risk Assessment and Management, 9(1/2).

Tilden, John. (2007). Personal Communication. In J. Spink (Ed.). East Lansing,
Michigan.

Todd, Ewen. (2008). Discussion of Counterfeit Risk Model. In J. Spink (Ed.). East
Lansing, Michigan: John Spink.

Town, Stephen. (2001). Crime displacement: the perception, problems, evidence and
supporting theory, Bradford District Architectural Liaison Oflice (UK).

Tritter, Donna L. (1989). EEC's Attempts to Stop the Importation of Counterfeit Goods,
The 12 B. C. Int'l & Comp. L. Rev, 423

Union des F abricants (on INTERPOL.int). (2004). Counterfeiting & Organized Crime
Report, . Union des Fabricants, 16, rue de la F aisanderie, 75116 Paris,: Union des
F abricants.

United Nations Development Programme. (1999). Human Development Report I 999.
New York Oxford, Oxford University Press: United Nations Development

Programme (UNDP).

United Nations International Narcotics Control Board. (2006). International Narcotics
Control Board Report: International Narcotics Control Board, United Nations

(UN/ NCB).

United States Department of Justice. (2006). Prosecuting Intellectual Property Crime
[Electronic Version]. CCIPS, Criminal Division, 2007. Retrieved September
2006 from USDOJ.gov,

<httD://www.l_rsdoi.gov/criminaI/cmercrime/imrganual/O1imapgb.

United States General Accounting Office. (1995). GAO Report to Congressional
Committees, FRUIT JUICE ADULTERATION: Detection Is Difficult, and
Enhanced Efforts Would Be Costly. (November 1995), GAO/RCED-96-l 8.

 

United States Government Accountability Office. (2008a). INTELLECTUAL
PROPERTY, Federal Enforcement Has Generally Increased, but Assessing
Performance Could Strengthen Law Enforcement Efforts [Electronic Version],
March from http://www.gao.gov/new.items/d08157.pdf.

United States Government Accountability Office. (2008b). RISK MANAGEMENT -
Strengthening the Use of Risk Management Principles in Homeland Security.
GA O-08-904 T.

United States Patent and Trademark Office. (2008). Laws and Rules - Code of Federal
Regulations and United States Code. Retrieved August 26, 2008, fiom
http://www.uspto.gov/main/glossary/lawsandrules.htrn

181

 

University of Florida. (2004). The Food Recall Manual. Gainsville, Fla: IF AS
Communication Services at the University of Florida, Food Science and Human
Nutrition Department.

US Customs and Border Patrol Bureau. (2008). National Workload Statistics [Electronic
Version]. Retrieved 6/20/2008 from
<http://www.cbp.gov/xp/cgov/about/accomplish/national_workload_stats.xml>.

US Department of Agriculture. (2005). Frequently Asked Questions, Revised [Electronic
Version]. Strategic Partnership Program Agroterrorism (SPPA) from
<http://www.usda. gov/documents/ SPPA-Q&A-1 0-05 .pdf>.

US Department of Commerce. (2007). Intellectual Property, What is It? Retrieved
February 5, 2007, from www.stopfakes.gov/sf_what.asp

US Department of State. (2006, January). Glossary of Intellectual Property Terms.
Retrieved December 15, 2007, fi'om
<http://usinfo.state.gov/products/pubs/intelprp/glossary.htm>

US Forest Service. (2006). Fire Danger Rating. WFAS, the Wildland Fire Assessment
System, USDA/US Forest Service Retrieved June 20, 2006, from
http://www.wfas.us/content/view/17/32/

Van Der Fels-Klerx, Ine H. J ., Goossens, Louis H. J., Saatkamp, Helmut W., & Horst,
Suzan H. S. (2002). Elicitation of Quantitative Data fiom a Heterogeneous Expert
Panel: Formal Process and Application in Animal Health. Risk Analysis, 22(1),
67-81.

van Schendel, Willem, & Abraham, Itty. (2005). Illicit Flows and Criminal Things:
States, Borders, and the Other Side of Globalization. Bloomington, IN.: Indiana
University.

Vladychenko, Alexander. (2006). Can the Council of Europe Support Actively the
Setting-up of Ejfective International Collaboration against Counterfeit
Medicines? Paper presented at the International Conference “ Combating
Counterfeit Drugs : Building Effective International Collaboration. fiom
http://mednet3.who.int/cfi/PRESENTATIONS/CouncilofEurope.ppt

Waite, Nick. (2004). The future of anti-counterfeiting, brand protection and security
packaging in Central and Eastern Europe. Leatherhead, Surrey, UK: Pira
International Ltd (PIRA).

Wallerstein, Mitchel B., Mogee, Mary Ellen, & Schoen, Roberta A. (Eds.). (1993).
Global dimensions of intellectual property rights in science and technology.
Washington, DC: National Academy Press.

Wang, J ingguo, Chaudhury, Aby, & H Raghav Rao. (2008). A Value-at-Risk Approach
to Information Security Investment, Information Systems Research. 19(1).

182

 

Wechsler, Jill. (2007). Industry, Regulators Move to Adopt Quality Standards.
Pharmaceutical Technology 31(2), 30.

Weisstein, Eric. (2003). CRC concise encyclopedia of mathematics (2nd ed.). Boca
Raton: Chapman & Hall/CRC.

WHO, World Health Organization. (2000). Counterfeit Drugs: Cottage Industry or
Organized Crime? Press Release.

WHO, World Health Organization. (2003). Assuring Food Safety and Quality:
Guidelines for Strengthening National Food Control Systems World Health
Organization

WHO, World Health Organization. (2004). Codex Alimentarius Commission,
PROCEDURAL MANUAL: World Health Organization

WHO, World Health Organization. (2006). The Safety of Medicines in Public Health i
Programmes: Pharmacovigilance an essential tool: World Health Organization. '

 

WHO, World Health Organization. (2007a). Counterfeit medicines. Retrieved March 1,
2007, 2007, from http://www.wpro.who.int/health_topics/counterfeit_medicines/

WHO, World Health Organization. (2007b). General information on counterfeit
medicines World Health Organization.

Willis, Henry H., & LaTourrette, Tom. (2008). Using Probabilistic Terrorism Risk
Modeling for Regulatory Benefit-Cost Analysis: Applications to the Western
Hemisphere Travel Initaitive in the Land Environment. Risk Analysis, 28(2), 25.

Withington, Neil. (2005). Second Global Congress on Combating Counterfeiting and
Piracy: Getting it right in the fight against fakes (Vol. Speech by: Director, Legal
& Security and General Counsel British American Tobacco).

World Customs Organization. (2003). Risk Management Guide.

World Customs Organization. (2005). Guidelines on controlling Free Zones in relation to
Intellectual Property Rights infringements.

World Customs Organization. (2006). The first review: Customs and Counterfeiting
2004, Press Release, [Electronic Version]. World Customs Organization (WCO),
2007. Retrieved August 18, 2008 from
<http://www.wcoomd.org[press/?v=1 &lid=1 &cid=1 89id=125>.

World Customs Organization. (2007a). Counterfeiting and Piracy - Crime of the 21 st
century? WCO NEWS, World Customs Organization (WCO)(54).

World Customs Organization. (2007b). Review of the Fight against Counterfeiting in
2006. Brussels, Belgium: World Customs Organization.

183

World Customs Organization. (2007c). SAFE - Framework of Standards: World Customs
Organization.

World Intellectual Property Organization. (2008). Are there internationally accepted
definitions of counterfeiting and piracy? Frequently Asked Questions Retrieved
November 18, 2008, from
http://www.wipo.int/enforcement/en/faq/counterfeiting/faq01 .htrnl,

World Trade Organization. (1994). Agreement on Trade Related Aspects of Intellectual
Property Rights, Including Trade in Counterfeit Goods. Retrieved May 4, 2007,
fiom http://www.wto.org/english/docs_e/legal_e/27-trips_05_e.htm

Xin, Hao, & Stone, Richard. (2008). TAINTED MILK SCANDAL: Chinese Probe
Unmasks High-Tech Adulteration With Melamine. Science, 322(5906), 1310-
1 3 l 1 .

Xuemei, Bian, & Cleopatra, Veloutsou. (2007). Consumers' attitudes regarding non-
deceptive counterfeit brands in the UK and China. Journal of Brand Management,
14(3), 211.

Yankus, Wyatt. (2006). Counterfeit Drugs: Coming to a Pharmacy Near You [Electronic
Version]. The American Council on Science and Health (A CSI-D. Retrieved
August 2006 from
<http://www.acsh.org/docLib/20060825_CDrqu_condensed.pdf>.

Yar, Majid. (2005). A Deadly Faith in Fakes: Trademark Theft and the Global Trade in
Counterfeit Automotive Components [Electronic Version]. Journal of
Criminology from www.intemetjoumalofcriminology.com.

Yeung, Godfiey, & Mok, Vincent. (2006). Regional monopoly and interregional and
intraregional competition: the parallel trade in Coca-Cola between Shanghai and
Hangzhou in China. Economic Geography, 82(1), 89.

Zaichkowsky, Judith Lynn. (2006). The psychology behind trademark infiingement and
counterfeiting. Mahwah, N.J.: Lawrence Erlbaum.

Zambrano, Lyda, Sublette, Kerry, Duncan, Kathleen, & Thoma, Greg. (2007).
Probabilistic Reliability Modeling for Oil Exploration & Production (E&P)

Facilities in the Tallgrass Prairie Preserve [Electronic Version]. Risk Analysis, 27,
1323-1333. Retrieved October 2, 2008.

184