Biotech valuation, An investor's guide - ING, 2004 - Cambridge ...

fettlepluckyBiotechnology

Dec 1, 2012 (4 years and 6 months ago)

586 views

ii
ii
i
l
EQUITY MARKETS
PharmaceuticalsWestern Europe
Dr Sally Bennett
(44 20) 7767 5852
sally.bennett@uk.ing.comRichard Parkes PhD
(44 20) 7767 5974
richard.parkes@uk.ing.com
Max Herrmann
(44 20) 7767 5873
max.herrmann@uk.ing.com
Biotech valuationAn investor’s guide December 2004
PLEASE SEE THE IMPORTANT DISCLAIMER, COMPANY DISCLOSURES AND ANALYST CERTIFICATION ON THE LAST PAGE OF THIS REPORT
ING Financial MarketsAMSTERDAMTel: 31 20 563 87 98
BRUSSELSTel: 32 2 557 10 26
LONDONTel: 44 20 7767 1000
NEW YORKTel: 1 646 424 6000
SINGAPORETel: 65 6535 3688
BratislavaTel: 421 2 5934 61 11BucharestTel: 40 21 222 1600BudapestTel: 36 1 268 0140Buenos AiresTel: 54 11 4310 4700DublinTel: 353 1 638 4000
EdinburghTel: 44 131 527 3000GenevaTel: 41 22 593 8050Hong KongTel: 852 2848 8488IstanbulTel: 90 212 258 8770KievTel: 380 44 230 3030
MadridTel:34 91 789 0030ManilaTel: 632 840 8888Mexico CityTel: 52 55 5258 2000MilanTel: 39 02 89629 3660MoscowTel: 7095 755 5400
ParisTel: 33 1 56 39 31 41PragueTel: 420 2 5747 1111SantiagoTel: 562 452 2700Sao PauloTel: 55 11 4504 6000SeoulTel: 822 317 1500
ShanghaiTel: 86 21 6841 3355SofiaTel: 359 2 917 6400TaipeiTel: 886 2 2734 7500TokyoTel: 813 5210 1500WarsawTel: 48 22 820 5018
Important Company Disclosures
The following designations [a-i] next to a company covered in this publication highlight that one or more
members of ING Group:
[a]holds 1% or more of the equity shares of the company (as at the end of the month preceding this publication).
[b]has lead managed or co-lead managed a public offering of the securities of the company in the last 12 months.
[c]has received compensation for investment banking services from the company within the last 12 months.
[
d
]
expects to receive or intends to seek compensation for investment banking services from the company in the
next 3 months.
[e]makes a market in the company’s securities in the US via ING Financial Markets LLC.
[
f
]
is a liquidity provider, or acts as designated sponsor or market maker, for the company on a German, French or
Dutch stock exchange.
[g]has lead managed or co-lead managed a public offering of the company's securities in the last 5 years.
[
h
]
has a member of its board of directors or su
p
ervisor
y
board or senior officer on the com
p
an
y
’s board of directors
or supervisory board.
[i] holds a net short position of 1% or more of the share capital, calculated in accordance with German law.
In addition, ING Group trades in the shares of the company/ies covered in this publication.Disclosures [a]-[e] are required specifically but not exclusively by US regulations.Analyst Certification and Disclaimer
The views ex
p
ressed in this re
p
ort accuratel
y
reflect the
p
ersonal views of the anal
y
st
(
s
)
about the sub
j
ect
securities or issuers and no part of the compensation of the analyst(s) was, is, or will be directly or indirectly
related to the inclusion of specific recommendations or views in this report.This publication has been prepared on behalf of ING (being for this purpose the wholesale and investment banking business of ING Bank NV and
certain of its subsidiary companies) solely for the information of its clients. ING forms part of ING Group (being for this purpose ING Groep NV and its
subsidiary and affiliated companies). It is not investment advice or an offer or solicitation for the purchase or sale of any financial instrument. While
reasonable care has been taken to ensure that the information contained herein is not untrue or misleading at the time of publication, ING makes no
representation that it is accurate or complete. The information contained herein is subject to change without notice. ING Group and any of its officers,
employees, related and discretionary accounts may, to the extent not disclosed above and to the extent permitted by law, have long or short positions
or may otherwise be interested in any transactions or investments (including derivatives) referred to in this publication. In addition, ING Group may
provide banking, insurance or asset management services for, or solicit such business from, any company referred to in this publication. Neither ING
nor any of its officers or employees accepts any liability for any direct or consequential loss arising from any use of this publication or its contents.
Copyright and database rights protection exists in this publication and it may not be reproduced, distributed or published by any person for any
purpose without the prior express consent of ING. All rights are reserved.
Any investments referred to herein may involve significant risk, are not necessarily available in all jurisdictions, may be illiquid and may not be
suitable for all investors. The value of, or income from, any investments referred to herein may fluctuate and/or be affected by changes in exchange
rates. Past performance is not indicative of future results. Investors should make their own investment decisions without relying on this publication.
Only investors with sufficient knowledge and experience in financial matters to evaluate the merits and risks should consider an investment in any
issuer or market discussed herein and other persons should not take any action on the basis of this publication. This publication is issued:1) in the
United Kingdom only to persons described in Articles 19, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order
2001 and is not intended to be distributed, directly or indirectly, to any other class of persons (including private investors); 2) in Italy only to persons
described in Article No. 31 of Consob Regulation No. 11522/98.Clients should contact analysts at, and execute transactions through, an ING entity
in their home jurisdiction unless governing law permits otherwise.
ING Bank N.V., London branch is authorised by the Dutch Central Bank and regulated by the Financial Services Authority for the conduct of UK
business. It is incorporated in the Netherlands and its London branch is registered in the UK (number BR000341) at 60 London Wall, London EC2M
5TQ. ING Financial Markets LLC, which is a member of the NYSE, NASD and SIPC and part of ING, has accepted responsibility for the distribution
of this report in the United States under applicable requirements.This document is classified as UK Inde
p
endent Research as defined in the ING Bank London Branch Research
p
olic
y
drawn up in accordance with FSA Rule COB 7.16.5R (2).
EQ_UK IND
Additional information is available on request
See back of report for important disclosures and disclaimer
Biotech valuation December 2004
Contents
Introduction to valuing biotech stocks 1
Overview of methodologies......................................................................................2
Fair value and trading value.....................................................................................3
Identifying a drug candidate 4
Cash-burning product companies 11
Introduction.............................................................................................................11
Overview of methods..............................................................................................11
Sales forecasting....................................................................................................21
How it works in reality: Target prices......................................................................24
Conclusions............................................................................................................27
Sensitivity analysis, scenario analysis and NPV evolution 28
Scenario analysis....................................................................................................28
Example 1...............................................................................................................28
Example 2...............................................................................................................29
Conclusion..............................................................................................................30
Profitable companies 31
Overview of techniques..........................................................................................31
Emerging profitable companies 33
Example 1...............................................................................................................33
Platform technology and tools companies 36
ING’s product NPV calculator 41
Appendix I 42
Ready reckoner NPV tables...................................................................................42
Appendix II 45
Rule of thumb metrics.............................................................................................45
Dr Sally Bennett
London
(44 20) 7767 5852
sally.bennett@uk.ing.com
Richard Parkes PhD
London
(44 20) 7767 5974
richard.parkes@uk.ing.com
Max Herrmann
London
(44 20) 7767 5873
max.herrmann@uk.ing.com
3 December 2004
Cover image courtesy of gettyimages
See back of report for important disclosures and disclaimer
1
Biotech valuation December 2004
Introduction to valuing biotech
stocks
In general, the valuation of biotechnology companies has been a
complicated business for investors.
Analysis of financial history gives little indication of future performance and non-risk-
adjusted forecast earnings growth underestimates the high risk of failure inherent in
drug development. In addition, complex science and highly speculative forecasting
make valuation a daunting task.
To add further complexity to the valuation dilemma, recent years have seen the sector
mature to the point that a sizeable set of hybrid biotech/pharmaceutical companies
have emerged, including emerging fully integrated pharmaceutical companies
(FIPCOs), specialty pharmaceutical companies and platform technology companies, to
which a broader range of valuation metrics (including more traditional ones) can be
applied. The investor is therefore presented with a diverse range of companies
encompassing cash-burning biotech companies that are many years from profitability
(if indeed they ever achieve it), profitable platform companies that operate
predominantly via a service model, niche pharmaceutical companies that are on the
cusp of profitability and hybrids of all three. Clearly, it is desirable to apply the most
relevant valuation metrics in each case in order to arrive at the most feasible and
informative value range for the company in question. However, the investor then faces
the danger that if different metrics are applied on an ad hoc basis simply according to
what ‘feels’ best, there is little comparability of results across the sector.
This guide therefore aims to outline the valuation tools that are available to investors
to assess the value of companies’ products and technologies and, to the greatest
extent possible, select the most appropriate valuation techniques while still allowing
for useful comparability across the sub-sectors. It is important to note that, although
the valuation methodologies are presented principally as a guide for investors, it is
hoped that company managements will also find them useful for appraising the value
of their own companies, as well as the values of competitors or potential acquisition
targets. It is, however, assumed that the investor is largely familiar with concepts
such as DCF and NPV.
It must nevertheless be made clear at the outset that, given the broad diversity of
companies operating in the ‘biotechnology’ and related life sciences arena, there is
no ‘correct’ way to value individual companies. In addition, ING has developed
metrics that, in our experience, offer some degree of predictive utility for correlating
the actual share prices of companies with their theoretical fair values. Overall, we
hope to be able to provide quantitative guidance to calculate a company’s valuation,
which will then be supported by qualitative share selection techniques. It should
always be borne in mind, however, that the valuation of a biotech company is
definitely still part art and part science.
See back of report for important disclosures and disclaimer
2
Biotech valuation December 2004
Overview of methodologies
The key metrics used in the valuation of biotech companies are as follows:

Comparable company analysis (based on the ratios of technology values, sales or
earnings to company enterprise values), usually abbreviated to ‘Comps’ or ‘CoCos’;

Standard PE or PEG ratios (where the company has a history of earnings);

Product NPVs, derived from risk-adjusted discounted future cash flows using a
terminal multiple; and

Discounted future earnings.
In practice, there are several variants and refinements of each of these, although they
share the same theoretical foundation.
In addition, one can employ more sophisticated tools, including ‘Monte Carlo’ analysis
and ‘real options’ analysis, which we touch on in later chapters. However, these methods
tend to be used more by company managements in valuing specific products or R&D
portfolios, rather than whole companies. In this regard, they can be useful for company
executives in understanding the risks and value-drivers relating to the development of
individual products, and can therefore help companies in managing their R&D portfolios
and product development and licensing activities. Despite this, they are rarely suitable for
investors in valuing whole companies for the simple reasons that:
a) They are typically time consuming and complicated;
b) In certain cases, they are not well validated; and
c) Most investors do not use them, so the results they give generally do not relate to
the reality of company share pricing.
Figure 1 highlights which of the commonly used valuation techniques are best suited to
different business models.
Fig 1 Biotech valuation methodologies
Complexity of
valuation technique
Biotech business
models
Emerging FIPCOs
Devices, diagnostics
Drug delivery
Platform technology
companies
Later stage product companies
Earlier stage product companies
C
o
m
p
a
r
a
b
l
e

t
e
c
h
n
o
l
o
g
y

v
a
l
u
e
(
u
s
e
d

t
o

a

g
r
e
a
t
e
r

o
r

l
e
s
s
e
r

e
x
t
e
n
t
)
Product NPVs
Discounted earnings –
where profitability is
visible in near term
Sales – where sales of
technology or products
are important but
profitability not yet
achieved
EV/sales
Earnings – once a
company has established
a track record of
profitability and growth
P/Es
PEG ratios
EV/EBITDA
EV/EBIT
Profitable platform
technologies
Specialty pharma
Increasingly
immature business
models
Source: ING
_
See back of report for important disclosures and disclaimer
3
Biotech valuation December 2004
Fair value and trading value
A common criticism of valuation methodologies, such as DCF, is that they are
theoretical, whereas the market prices of companies (or company shares) are actual,
quantifiable measures of market supply and demand, which almost never match the
fair values ascribed by analysts. While this is true for all quoted companies, the
argument runs that in view of biotech’s speculative nature and high volatility, it is
particularly true of this sector. Nevertheless, ING regularly monitors the variance of
company share prices from their respective NPVs (nearly always a discount) across
the European Biotech sector and applies adjustment factors based on an average
sector discount to arrive at what might be more realistic trading values for companies.
This is described in more detail later in this report.
See back of report for important disclosures and disclaimer
4
Biotech valuation December 2004
Identifying a drug candidate
Central to understanding how to value biotech companies is an appreciation of the
drug discovery and development process. The high costs and long time-lines involved
in this process form the basis of the high risk/reward profile of most biotech
companies. In addition, although the drug discovery and development process is long
and complex, biotech companies can add value along the value chain in different ways
by providing services or technologies that facilitate and speed up one or more process,
or ultimately by developing their own drugs. Understanding the value that companies
capture and the risks they incur in doing so is key to evaluating their worth.
Fig 2 Drug development timelines
Compound discovery and development Clinical trial phase
M
A
R
K
E
What accomplished
Target
identification
Target
validation
T
Approval for
clinical trial
Lead
compound
identification
Lead
optimisation
Obtain
toxicology and
pharmacokinetic
data
Ascertain
optimal dose
level, dosing
regimen, route
of administration
indication, establish
appropriate
endpoint for
phase III studies
Demonstrate
statistically
significant
effect of drug
Obtain human
pharmacokinetic
data, safety
data
Approval
for matching
FDA review
61 2 3 4 5 II IIII
Early discovery
How accomplished
Expected timeframe
Probability of approval
Genomics
traditional
biomedical
research
Animal
models,
in vitro
assays
Assay
development.
Rational drug
design. High
throughput
screening,
combinatorial
chemistry
Medicinal
chemistry
In vivo assays,
animal models
Patients Patients
Discovery: 2-3 years
12-18 mths 30 days 1-2 years
30%
1-2 years
70%
½-1 year
90%
Healthy human
volunteers
½-1 year
10%
Attrition rate
10,000 compounds 250
10
1 compound
Traditional Drug Development
What accomplished
Source: ING
Drug discovery is a complex process that offers enormous opportunities to
biotechnology companies. At almost every step along the drug development pathway,
improvements are possible and entirely new technologies are continually being
developed and applied to enhance what was once a relatively archaic process.
The first step in identifying a new drug for a particular indication requires identifying a
relevant protein, or other molecule, whose function can be modulated to affect the
course of disease. That is done in the first stages of the drug discovery process, also
called target identification. Next, one has to be absolutely sure that this target has a
beneficial effect on the disease, which is tested during a process known as target
validation. Now the hunt is on for a molecule that interacts with this newly discovered
target. For traditional drug development, the identification of a target for intervention is
simply the starting point against which a potential drug candidate can be selected
through a process known as lead compound identification and optimisation.
See back of report for important disclosures and disclaimer
5
Biotech valuation December 2004
What is a target?
Most drug targets in the cell are proteins. Proteins are the machinery of the cell and
come in many forms. Proteins can be cell receptors that receive chemical signals from
other cells and set in motion a cascade of processes, also called pathways, finally
resulting in the change in the cell demanded by the original signalling chemical
(ligand). If the ligand (for example, an epidermal growth factor (EGF) molecule) gives a
signal to cancer cells to start repairing and replicating themselves, a treatment could
be devised to shut down the receptor type that binds the ligand (in this case known as
the EGF receptor, or EGFr), giving cancer cells no possible escape route if they are
damaged by chemotherapy or radiotherapy. That could be done by blocking the EGFr
in such a way that EGF cannot bind to it, thereby blocking the entire subsequent
cascade process and hence the final outcome. Or, one could block EGF instead, by
devising a small molecule that fits perfectly in the three-dimensional shape of the
signal, leading to a change in the configuration of the EGF protein and thus making it
impossible for the signal to bind to the receptor.
Proteins perform different functions in the cell
Other proteins fulfil the role of catalysts. Just like catalysts in cars, they are necessary
to facilitate a chemical or biological reaction. In the absence of the catalysts, also
called enzymes, biological reactions would not take place, or would only occur at a
very slow rate. Large families of enzymes exist and they have different family names
such as kinases, proteases, etc. One family can have a thousand related and
structurally similar enzymes. So the protein encoding regions of genes discovered in
sequencing are assigned to gene families or classes based on amino acid homologies
and motifs. For example, all proteins exhibiting homology to motifs present in G-protein
coupled receptors (GPCR) are grouped into one class. Classes that are known to be
‘druggable’ (that is, classes of proteins that in the past have been successfully
modulated by small molecule drugs) are of most interest. If a cancer cell relies heavily
on a specific enzyme to grow or to replicate, a possible drug could act to inhibit that
enzyme, thereby preventing the cancer cell from replicating and causing it to die.
Most biotechnology companies operate by elucidating drug targets, linking them to
pathways in the cell and developing novel treatments directed at ameliorating
deficiencies in these pathways.
To date, the entire pharmacopoeia of modern medicine mediates its action through
one or more of an estimated 483 different targets, consisting of a mix of receptors,
enzymes, hormones, ion channels, DNA and nuclear receptors (see below). Increasing
pressure on the pharmaceutical industry to produce innovative products is however
driving research into identifying drugs to novel drug targets. This process will
principally be driven through research emerging following the advent of genomics and
the subsequent sequencing of the human genome. It is estimated that the number of
drug targets available to the pharmaceutical industry is set to increase to as many as
4,000 druggable targets.
Each of these targets must first be identified and later validated as a target for
intervention. This process typically takes two to three years. For the new biotechs, this
is typically a protein whose presence or absence is associated with a disease. The
protein, or another that interacts with it, is then tested for pharmacological effect.
See back of report for important disclosures and disclaimer
6
Biotech valuation December 2004
Fig 3 Drug targets currently utilised by the pharmaceutical industry
Receptors
45%
Hormones and
factors
11%
DNA
2%
Nuclear receptors
2%
Unknown
7%
Ion Channels
5%
Enzymes
28%
Source: Drews, J., Science Vol 287 p1960
_
Lead identification
Once a target for drug intervention has been identified, a candidate molecule must be
isolated that modulates its function and thus the selected progression of a disease.
For small-molecule drugs, this can involve screening thousands of compounds in an
assay that measures activity. This process has been accelerated by the development
of combinatorial chemistry (automated synthesis of a ‘library’ of similar compounds, in
order to find the most useful) and through high-throughput screening (automated
testing of thousands of potential compounds from a ‘library’). Antibody therapeutics
can similarly be developed using libraries of molecules (phage display) or through
immunisation of animals.
Many biotech companies develop products in which the target acts as a drug
itself. These would include protein therapeutics such as Amgen’s erythropoeitin
(EPO) and genes utilised in gene therapy products. Development of these
products bypasses the requirement for expensive and lengthy lead optimisation.
Antibody therapeutics also require shorter time in lead optimisation. Although
selection of antibodies with the required specificity and affinity requires some
manipulation, all antibodies are based on naturally occurring proteins based on
a limited number of backbones. Their absorption, distribution, metabolism and
toxicological profiles are therefore well defined. Individual antibodies vary only in
their specificity and in some cases circulating half-life. What’s more, in contrast
to the large cost of maintaining teams of highly skilled medicinal chemists,
which remain the domain of large pharmaceutical companies, the skills needed
to develop protein and antibody therapeutics are available to, and reside for the
most part in, biotechnology companies.
Lead optimisation
Once a drug candidate has been identified with the required activity and specificity
further refinement is needed before it is ready to begin enter clinical trials. In its current
state a lead is often not fit to become a good drug. It should be polished, worked on,
etc. Sometimes the chemical structure of the ‘hit’ is not right, meaning that it is not very
well adsorbed in the bloodstream, or it is metabolised in the cell too rapidly, or it does
not fit perfectly with the target. So, chemists will adapt the structure to give it more
See back of report for important disclosures and disclaimer
7
Biotech valuation December 2004
drug-like properties, but without adapting it too much that it loses its ‘hit’ capacity (or
pharmacological properties). A balance has to be found, and every time the molecule
is improved it has to be tested again in a secondary screen. For small molecule drug
development this process principally requires the creation and optimisation of
compounds with appropriate absorption, distribution, metabolism and elimination
profiles – a process known in medicinal chemistry as ADME. This step remains a
considerable bottleneck in traditional drug development. The resolution of medicinal
chemistry problems remains formidable and typically requires large teams of expert
chemists and biologists.
Small molecules versus biologicals
In simple terms, drugs act by binding to a target in a way that modulates its
function. Small molecules that bind very well to one specific target only are
traditionally the focus of pharmaceutical companies. Small molecules have the
great benefit that they are taken orally and can pass the gastrointestinal system
to be absorbed in the blood stream. Taking drugs in pill format is a major
advantage and, as such, many of the blockbuster drugs that treat chronic
conditions, such as depression (Prozac), heart disease (Lipitor) or allergies
(Zyrtec), are based on small molecules.
In contrast, biologicals, which include proteins and antibodies, are
macromolecules that are either broken down in the gut or cannot pass through
the intestinal wall and therefore cannot be administrated orally. Due to the
requirement for delivery by injection the market potential of biologicals can be to
some degree limited. Despite this, as most under development are innovative
products addressing areas of unmet medical need, premium pricing is
obtainable. A factor highlighted by the phenomenal success of erythropoeitin,
which now has worldwide sales of over US$4bn.
Pre-clinical testing: includes lab assays and animal models, to establish its specificity
for the target, toxicity in various does and pharmacokinetics (changes in concentration
over time). Pre-clinical trials typically take up to one year.
Investors should not place too much emphasis on pre-clinical data. By
definition, most of these drugs will go on to fail in the clinic, so we believe that
investors should view pre-clinical data more as a proof of concept of the
mechanism of action of a new drug rather than as an absolute measure of the
drug’s efficacy. In summary, we place no specific value on drugs that have only
generated pre-clinical data.
Clinical development
Once a drug candidate has been selected, results of pre-clinical testing are submitted
to the regulatory authorities – EMEA (European Medicines Evaluation Agency) or, in
the US, the FDA (Food and Drug Administration) – in the form of an IND
(investigational new drug application). If these authorities deem that pre-clinical
evidence justifies human testing then clinical trials can begin.
See back of report for important disclosures and disclaimer
8
Biotech valuation December 2004
Prior to entering clinical development, meticulous attention must be paid to the
planning of trials in order to optimise a drug’s chance of success. This design of the
trial protocol must identify:

The purpose and objectives of the trial;

The entry and exclusion criteria for patients to be enrolled in the trial;

Clinical endpoints; and

The statistical plan, which will ensure that the drug can be shown to have
statistically significant effect on disease.
The clinical trial process is divided into three stages known as Phase I, II and III.
Phase I. Phase I trials are conducted on healthy volunteers (generally 20-80
individuals) and aim to establish side effects and the pharmacokinetics profile.
Interpretation of Phase I data primarily focuses on any incidences of so-called adverse
events, such as nausea, vomiting, headache, etc. Other measures of toxicity will
include the effect on organ function, blood enzyme levels and any other indicators of
physiological and biochemical disturbance. The prevalence of a particular side effect,
which will be tolerated by the regulators, depends on the severity of the disease to be
treated. As a result, lifestyle drugs such as treatments for irritable bowel syndrome
must have very clean side-effect profiles – GlaxoSmithKline famously had to withdraw
its IBS treatment, Lotronex, due to incidences of a potentially life threatening syndrome
following marketing of the drug. In contrast, for life threatening conditions such as
cancer, side effects are generally more acceptable.
In some high-priority, life-threatening conditions, such as Aids and cancer,
developers are allowed to speed up the process by testing on patients from the
outset – such trials are designated Phase I/II trials (I for safety, II for patients).
Phase II. Phase II trials are conducted on a few hundred patients and typically involve
comparison with a control group. As such, these trials often provide the first evidence
of efficacy of a drug and are thus termed as ‘proof of principle’ clinical trials. They also
aim to provide information for optimisation of the dosing and administration of the drug
as well as helping to identify which patients are most likely to receive benefit from
receiving a particular drug. The small patient size tested in Phase II trials normally
means that only a trend towards evidence of efficacy can be shown as results are of
limited statistical significance.
Phase III. Phase III, or ‘pivotal’ trials, aim to provide statistically significant evidence of
efficacy and safety in larger patient groups. Often carried out at multiple trial sites,
Phase III must provide additional safety information and, crucially, determine whether
the drug is inferior, equivalent, or superior to existing treatments.
See back of report for important disclosures and disclaimer
9
Biotech valuation December 2004
The p-value is a measure of the statistical significance of a trial outcome. It
gives the probability that a particular outcome is caused by coincidence, rather
than the effect of the drug. A p-value of 0.05 indicates that there is a 5%
probability that this result was merely a chance occurrence. The FDA and
EMEA will accept as statistically significant a p-value that is less than 0.05
although, as part of its review process the FDA re-analyses the data under more
stringent criteria and can therefore recalculate a much higher p-value.
Therefore, more confidence should be possible where the p-values are 0.01 or
less. The most important p-value is that associated with the primary end point of
the trial. In Phase III, the primary end point is the one that determines whether a
drug is approved or not and looks at a definitive outcome of disease, eg, death,
progression to disability, rate of heart attack), while the secondary end points
are used to support the data and look at symptomatic parameters that are
related to the intensity of disease (eg, quality of life data).
Approval and market. After completion of trials and analysis of the results, assuming
a positive outcome, a vast submission is filed for regulatory review (a recent
application by Amgen for a rheumatoid arthritis treatment ran to over 500,000 pages).
This, in particular, is a minefield, since the package submitted to the regulatory bodies
has to stand up to the deepest level of scrutiny by leading experts in the therapeutic
area (the regulatory panel). These experts decide whether a drug should be approved
and under what circumstances it will recommend use of the drug. These
recommendations make up the ‘label’ specifying patient types and circumstances.
After receiving approval, further ‘Phase IV’ studies are often undertaken that gather
additional data on efficacy and costs and benefits in ‘real-life’ conditions of extended
treatment and comparisons to other marketed treatments.
The FDA approval process generally takes more than a year to complete. Although
reforms have resulted in a dramatic decrease in this time, from a mean of 32.6 months
in 1992, to 15 months in 2000, approval times have increased during the last two
years. These increases along with a number of high-profile drug failures and
withdrawals have led to fears that the authorities are tightening their controls on drug
approvals. However, in recent times there has been an increase in development of
drugs targeting ‘quality of life’ conditions – these include depression, anxiety, migraine
and irritable bowel syndrome. These have very large potential markets and therefore
attract large pharmaceutical companies searching for ‘blockbuster status’ drugs.
However, the non-life threatening nature of these disorders are likely to lead the FDA
to a more cautious stance in clarifying the safety profile of these drugs.
In addition, with an ageing Western population and an increase in the number of new
drugs in development and on the market, increased prescribing pressure on physicians
is likely to result in increases in unidentified drug interactions and adverse events. This
may result in more drugs being withdrawn over the next decade. The FDA’s post-
marketing surveillance system will come to play an ever more prominent position in the
monitoring of prescribed drugs.
See back of report for important disclosures and disclaimer
10
Biotech valuation December 2004
FDA approval time reforms
In recent years, attempts to streamline and accelerate the approval process
have been fuelled by consumer calls for speedier access to new drugs,
especially for life-threatening conditions, along with pressure from the
pharmaceutical industry. As the length of time of this process effectively eats
into the patent life of a new drug, therefore reducing any period of market
exclusivity, a reduction in the time taken for drug approvals is of considerable
benefit to the pharmaceutical and biotechnology industry.
Improvements in the approval process have principally been instigated through
reforms precipitated by the Prescription Drug User Fee Act of 1992 (PDUFA) and the
FDA Modernisation Act of 1997. These have included the instigation of ‘fast track’
approval. A drug that is intended for the treatment of a serious or life-threatening
condition may be placed on a fast track if it demonstrates the potential to address
unmet medical needs for the condition. The fast track approval process cuts the time
for approval from the usual +12 months to 6 months. Critics, fuelled by withdrawals of
drugs, such as Lotronex, have accused the FDA of compromising safety by initiating
mechanisms to speed approval. Nevertheless, according to the Tufts Centre for the
Study of Drug Development, the rate of drug withdrawals due to post-marketing safety
issues has, in fact, remained stable.
Product lifecycle
Pharmaceutical products have lifecycles in the same way that most other
commercialised products do. Figure 4 shows a rough schematic of a theoretical
blockbuster product’s sales following launch, including an initial uptake phase, a
growth phase, a mature phase and a sharp drop in sales following patent expiry in year
ten. Clearly, some products behave differently, realising slow or linear growth, sales
declines due to competition before the advent of patent expiry, or in certain cases a
less marked drop following patent expiry and generic competition. However, the
diagram does serve as a useful approximation of the lifecycle of many drugs.
Fig 4 Product lifecycle
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Years post launch
Sales (US$m)
Introductory
phase
Growth phase
Mature phase
Post patent
expiry phase
Source: ING estimates
_
See back of report for important disclosures and disclaimer
11
Biotech valuation December 2004
Cash-burning product
companies
Introduction
So-called ‘cash-burning product companies’ probably make up the majority of ING’s
European biotech universe and remain a central focus of investor interest. These are
companies with little or no revenue that invest heavily in developing drug pipelines, with
the aim of licensing and eventually commercialising the products. These companies’
pipelines usually comprise of one or more pre-clinical or clinical stage drug, in which the
majority of the companies’ value lies. Occasionally, some of these companies will have a
marketed product, usually of relatively modest sales potential (such as Ark Therapeutics’
wound dressing Kerraboot, or Medigene’s in-licensed prostate cancer treatment Eligard),
or a small royalty interest in a major product, such as Cambridge Antibody Technology’s
rheumatoid arthritis treatment, Humira. Nevertheless, overall, the revenues generated by
these products do not cover the companies’ major R&D expenditures and they remain
net cash burning. Our preferred valuation method for these companies uses a risk-
adjusted, product-based NPV analysis.
Valuation of classical drug discovery biotechnology companies is a notoriously difficult
task. Analysis of financial history gives little indication of future performance and non-
risk-adjusted forecast earnings growth underestimates the high risk of failure inherent
in drug development. Despite this, for products currently in development – or indeed
licensing deals that encompass product development – sales potentials are easier to
forecast, and this can provide the starting point for valuing those products. In addition,
clinical development risks, measured as product attrition failure rates incurred as drugs
progress through the clinic, are reasonably well documented and can be applied as
meaningful risk-adjusters to account for the chance of development failure. We have
therefore chosen as our preferred method to utilise a risk-adjusted product-by-product/
deal-by-deal-based NPV analysis.
Overview of methods
Most investors will be aware that the net present value is simply the value of a future
cash, revenue or earnings stream discounted back to today’s value according to
investors’ required rates of return. Most biotechnology valuations take the NPV as the
sum of a series of risk-adjusted discounted cash flows arising from the company’s
various clinical projects, assessed over a time horizon during which at least one of
these projects will yield a commercialised drug. However, there are various ways of
performing this exercise with varying degrees of complexity, depending on what the
investor is trying to achieve. For example, an investor trying to assess the values of
two or more broadly comparable companies is not necessarily going to want to expend
the same time and energy as someone appraising a single drug. Note that all DCF-
based methods use as their starting point an estimate of product sales. A brief
discussion on sales forecasting and how to arrive at peak sales figures is given later in
this report.
We utilise a risk-
adjusted product-by-
product/deal-by-deal-
based NPV analysis
See back of report for important disclosures and disclaimer
12
Biotech valuation December 2004
rNPV method
The rNPV (risk-adjusted NPV), otherwise known as ENPV (expected NPV) method, is
the most theoretically rigorous way to value a drug. This method explicitly considers all
of the cash flows and relevant costs resulting from a project, including sales or royalty
revenues, milestones, COGS, R&D costs, clinical trial costs, pre-marketing and SG&A
costs, and so on. The project is appraised over a 20- to 25-year time horizon,
encompassing the product’s full lifecycle from its current stage in clinical development,
registration, launch, growth, peak sales, patent expiry and post-patent sales. The
resulting cash flows are then risk-adjusted using clinical product attrition rates in order
to reflect the risk of the product failing in clinical trials, before being discounted at an
appropriate discount rate, typically around 13%.
Some practitioners advocate using a much higher discount rate than this, in the order
of 30-40% that is sometimes used by VCs. However, if product risk is taken fully and
correctly into account in valuations for biotechnology companies, there is no reason to
use abnormally high discount rates. A figure of 13% approximates to the long-term
trend in the total return from equities.
While theoretically rigorous and often useful, the rNPV method is heavily dependent on
a number of assumptions, most importantly regarding product sales and sales
evolution, and therefore requires a significant investment of time and resources. This
method tends to be more suitable for valuing individual drugs, for portfolio
management or licensing reasons, than whole companies. If a whole company is to be
valued, further assumptions need to be made regarding capex, changes in working
capital and non-product related expenditures/overhead. However, this process may be
a little heavy going for investors or managements looking to appraise a broader
universe of companies, particularly when several valid simplifying assumptions can be
made.
Fig 5 rNPV analysis of partnered product in Phase III development
Phases 2004 2005 2006 2007 2008 2009 2010 2011 2012 2024
Partner revenues 0 0 0 6 22 44 77 100 110 2.3
Royalty (25%) 0 0 0 1.5 5.5 11 19.3 25 27.5 0.6
Milestones 30
Phase III trial costs -5 -5
Registration costs -1.5
Overheads (%) 0 0 0 -0.5 -1.7 -3.3 -5.8 -7.5 -8.3 -0.2
Net cash flows -5 -5 28.5 1.1 3.9 7.7 13.5 17.5 19.3 0.4
Probability of reaching this point (%) 10
0
100 7
8
7
0
7
0
70 7
0
7
0
7
0
70
Risk-adjusted cash flows -5 -5 22.2 0.7 2.7 5.4 9.4 12.3 13.5 0.3
Discount factor (13%) 1 0.88 0.78 0.69 0.61 0.54 0.48 0.43 0.38 0.09
DCF -5 -4.4 17.4 0.5 1.7 2.9 4.5 5.2 5.1 0.0
NPV 48.3
Note: Years 2013-23 were omitted for space reasons. Patent expiry expected in 2016.
Source: ING estimates
_
See back of report for important disclosures and disclaimer
13
Biotech valuation December 2004
Why use risk adjustment?
As we noted earlier, drug development is a risky business and there is a strong
chance that any given product could fail to meet its clinical trial objectives and
hence never make it to market. Clearly, some form of adjustment needs to be
made to the valuation to reflect these additional tiers of risk. Nonetheless,
estimating the probability that a drug will make it to market is not purely a matter
of subjective judgement and experience. Various analyses have been done,
most notably by Joseph DiMasi at the Tufts Center for the Study of Drug
Development, to show the relative percentages of NCEs from pharmaceutical
pipelines that manage to progress from IND through the various stages of
clinical development to market. From these attrition values, a probability of
reaching the market can be inferred for a drug at any given stage of clinical
development. For example, Di Masi’s data shows that from 1987 to 1992 there
was a 23% chance of a drug in Phase I reaching the market, while for drugs in
Phase II and III, the probabilities were 33% and 79%, respectively.
Fig 6 Drug clinical success rates by Phase success rates
24.2
22.6
33
32.7
72.9
78.5
0
10
20
30
40
50
60
70
80
90
1981-1986 1987-1992
Percentage
Phase I
Phase II
Phase III
Source: Dimasi et al
_
Nevertheless, it is important to put these figures in context and to appreciate that they
can serve, at best, as a guideline. Not only do figures from different sources, such as
Tufts CSDD and the Centre for Medicines Research (CMR), differ from one another,
but when broken down into therapeutic sub-categories, different data shows that some
therapeutic areas, such as central nervous system (CNS) drugs, are historically much
riskier than others, such as anti-infectives. Similarly, biologics do not share the same
attrition profiles as small molecules, and biotechnology companies can often be riskier
propositions than larger pharma. We explain later in the chapter how to account for
these factors.
See back of report for important disclosures and disclaimer
14
Biotech valuation December 2004
Fig 7 Clinical success rates by indication
0 5 10 15 20 25 30
Anti-infective
Analgesic/anaesthetic
Gastrointestinal
Endocrine
Cardiovascular
Antineoplastic
Immunologic
Central nervous system
Respiratory
Miscellaneous
Success rate (%)
Source: Dimasi et al
_
Assessing the possibility of a drug succeeding – other
influences affecting the chances of success
Earlier we noted the principle of applying figures derived from historic product attrition
rates as risk-adjustment factors in valuing product NPVs. There are many sources of
these figures, and in most cases we apply industry-accepted standard rates, shown in
Figure 8. However, it is important to bear in mind the limitations of these figures, and
how they can be sensibly adjusted to reflect the reality of cash-burning biotech
companies better.
We use the following well-documented success rates for our biotech
universe
Fig 8 Clinical trial success rates (%)
Phases Phase to Phase Phase to market
Pre-clinical 50 5
Phase I 33 10
Phase II 43 30
Phase III 78 70
Registration 90 90
Market 100
Source: ING
_
The attrition figures are based on historical experience, mainly from large cap
pharmaceutical companies whose development portfolios are relatively large and have
traditionally comprised a rough 70:30 mix of ‘me-too’ products against well-validated
targets and innovative products against novel targets. There is some evidence that
smaller biotech companies, which have just a few, higher-risk products in the pipeline,
may use less restrictive guidelines to advance products into the next stage of
development. If the late-stage success rate correlates with the size of the pipeline, it
will be lower for smaller companies. In addition, novel targets have a lower success
rate because there is less information available about optimal formulation, dosing, side
effects and efficacy. Because of the pipeline shift towards novel targets, the success
rate for late-stage products may already have decreased from 70% to 50%
1
.

1
Roche Annual Meeting 2002
See back of report for important disclosures and disclaimer
15
Biotech valuation December 2004
Indeed, a cursory analysis of the fates of European biotech products undergoing
Phase III development since 2000 indicates that the true percentage of those products
successfully reaching the market may in fact be closer to 63%. Of the 32 products
documented in Phase III trials, 20 went on to be approved while 12 were rejected.
When specialty pharmas, such as Shire and Lundbeck, and drug-delivery companies
such as SkyePharma, are omitted, the figure decreases to only nine successes (43%),
which may represent a more realistic European biotech success rate.
Fig 9 Outcomes of drugs undergoing Phase II trials since 2000
Successes Drug Indication
Acambis Arilvax Yellow feve
r
Actelion Tacleer Pulmonary Arterial Hypertension
Biosearch Italia Ramoplanin Bacterial infection
Cambridge Antibody Technology Humira Rheumatoid arthritis
Celltech Mylotarg Acute leukaemia
Elan Antegren Multiple sclerosis
Lundbeck Cipralex Depression
Medigene Polyphenon E Cancer recurrence
Oxford Glycosciences Zavesca Gaucher’s disease
Photocure Metvix Basal cell carcinoma
Serono Rebi
f
Multiple sclerosis
Serono Raptiva Psoriasis
Serono Serostin Hormone supplement
Serono Gonal-F Ovulation induction
Shire Reminyl Alzheimer’s disease
Shire Fosrenol Hyperphosphatemia
SkyePharma Depodur Moderate-to-severe pain
SkyePharma Uroxatral Benign prostatic hyperplasia
SkyePharma Foradil Asthma
Vernalis Frova Migraine
Failures Drug Indication
Actelion Veletri Acute heart failure
Antisoma Pemtumomab Ovarian cance
r
Cambridge Antibody Technology Trabio Glaucoma surgery adjunct
Celltech Humicade Crohn’s disease
Celltech BMS 275291 HIV-Related Kaposi’s Sarcoma
Exonhit Therapeutics Ikomio Amyotrophic Lateral Sclerosis
Genmab HuMax CD4 Rheumatoid arthritis
Medivir MIV-606 Shingles
PPL Therapeutics AAT Fibrin sealant
SR Pharma SRL 172 Non-small cell lung cance
r
Xenova Tariquidar Drug resistance modulato
r
Zeltia Yondelis Cance
r
Source: Company data, ING
_
In any case, it is not unreasonable, when assessing product risk-adjustment rates, to
establish whether a product is an innovative one or a me-too product. In addition, the
disease category for which the drug is being developed is important. Drugs against
diseases whose underlying pathologic mechanism are well understood (such as
hypertension, diabetes, bacterial infection or asthma) have a higher probability of
success than do those for diseases for which there is little knowledge about the
molecular biology (such as sepsis, CNS or cancer).
See back of report for important disclosures and disclaimer
16
Biotech valuation December 2004
Standard NPV using terminal multiple
Although a thoroughly researched rNPV model can provide superior insight into the
value drivers of biotech companies, where it is not practicable to make such a
significant time investment, a ‘standardised’ NPV-based methodology can be
employed that produces valuations which are within reasonable proximity to the results
of rNPV methods, but which employ several simplifying assumptions that make the
whole process very much quicker. This method essentially averages several variables
while focusing on those to which the end valuation is most sensitive, such as peak
sales, discount rate and investment appraisal period. Furthermore, the methodology
does, on average, produce valuations that can be surprisingly accurate in determining
whether companies are overvalued or undervalued with respect to their peer group.
Firstly, we consider the case of companies that derive revenues (or will derive
revenues) exclusively on a partnered royalty basis (which is true of a large cross-
section of the European biotech universe). These companies do not generally incur
COGS or direct selling and marketing costs (for those companies that do retain
manufacturing rights, we assume the COGS are largely variable costs and are
covered, with negligible transfer price, within the royalty payments).
Once an estimate of peak sales has been made (see section on market forecasting), a
standard growth curve is applied, which assumes roughly sigmoidal sales grow to a
peak level, usually over five years. Royalty rates for each product, if not known
precisely, can be estimated according to an approximate schedule:

For products in Phase I or II – 15% (it is assumed a product in Phase I will not be
licensed until it reaches Phase II); and

For products from Phase III to market – 25%.
If there is clearly a case for doing so, an adjustment can be made to these royalties (or
to the actual royalty rates, if disclosed) to account for ‘royalty stack’, ie, the royalties
the biotech in turn owes previous licensors or related-IP holders with a claim to the
product (such as the payments made by antibody companies to Genentech under the
Cabilly patents). Milestone payments are added to royalties or sales to generate total
revenue payments (if no licensing deal has been signed or if milestones have not been
announced, milestone values can be estimated from previous benchmark licensing
deals).
An overhead of 30% is then deducted from the combined revenue stream to account
for R&D and G&A costs, and tax is also applied at 30%. Net cash flows are risk
adjusted using the same mechanism as that used in the rNPV method, and risk-
adjusted cash flows are then discounted as normal.
Where companies retain full rights to their products and market (or plan to market)
themselves, we deduct an additional percentage of revenues to account for COGS and
selling and marketing costs for those marketed products.
A key difference between the standard method and the rNPV method is that only the
first five years of product sales are considered. It is not therefore usually necessary to
assess the impact of patent expiry, as this will most likely occur after the time frame
being examined. This method also has the advantage that it negates the increasing
uncertainties associated with estimating free cash flows beyond a certain point in the
future. We would typically apply a terminal multiple of 20 times the product’s fifth-year
contribution to arrive at a terminal value and take the product’s value as the NPV of
See back of report for important disclosures and disclaimer
17
Biotech valuation December 2004
discounted free cash flows from year zero (ie, now) to year four of sales, plus the
terminal value.
Note that this approach will provide an NPV value per product that is typically greater
than that which would be arrived at by modelling each product’s full cost and revenue
profile using the rNPV method with no terminal multiple. A key assumption using the
standardised method, therefore, is that profits from the marketed product will be re-
invested into the business to generate further returns; ie, the product is valued as part
of an ongoing business, rather than as a standalone asset.
Clearly, it is expected that the company will continue to generate or add value to drugs
beyond those that are currently represented in its pipeline. Thus, comparisons
between biotechnology companies with important products expected to succeed
amount to a comparison between potential sales levels, margins, launch dates and
risks associated with key products.
Other methods: Discounted future earnings
We believe that the rNPV or standardised NPV methodologies are the most
theoretically rigorous and rational methods of valuation, as well as being independent
of investor sentiment. Some investors, particularly in the US, favour related methods,
most notably the discounted future earnings approach. This method can be used in
two ways:

If the company is on the cusp of profitability (ie, two years or less), an average
high-growth biotech PER multiple is applied to the forward earnings estimate of the
company to be valued; and

If the company is some years from profitability, a standard multiple is applied either
to the first or second year’s earnings, or an average of the two, which is then
discounted back to a present value using a 25-40% discount rate.
The discounted earnings method involves the construction of a future earnings model
for a biotechnology company, whereby analysts would typically include full revenues
for all products close to market (no earlier than Phases II and III). A forecast profit and
loss account is then constructed from these assumed revenues and an average PER
multiple for growth (profitable) biotechs is applied. In the first scenario, the multiple can
be applied to one of the company’s first few years of profitability (a lower multiple might
be used for year two than for year one, as shown in Figure 10, to reflect the gearing
attributed to the high growth in earnings.
Fig 10 Discounted future earnings model
No. of years to profitability
0 1 2
Typical multiple applied (x) 60 -80 40-45 25-30
See historic valuations applied to companies such as Actelion, Celgene, CV Therapeutics, Gilead, ImClone, MGI,
Pharmion and Neurocrine in their first years of profitability as a guide to these multiples.
Source: ING
_
In the second case, where the biotech is still some years from profitability, a more
standard multiple of 35-45x is applied to an average of the first two years of profitability
and that number is discounted back to a fair value for today. A discount rate is used to
reflect the required rate of return of a high-risk investment (typically 25-40% pa).
Both methods, in particular the second, have limitations. It could be argued that the
discounted future earnings method does have the advantage that it tends to rely on
See back of report for important disclosures and disclaimer
18
Biotech valuation December 2004
fewer assumptions than the NPV methods (for example, assumptions regarding capex
are not necessary), notwithstanding that the value depends more heavily on those
assumptions that do have to be made. Equally, forecasting P&L figures, including
product revenues, two years into the future can be far more accurately done than
forecasting figures several years into the future.
However, basing a valuation on earnings leads inevitably to a loss of comparability
between pan-European companies because of different accounting conventions, tax
regimes and capital structures. In addition, applying a multiple essentially renders
the results a relative measure, rather than an absolute one. Furthermore, an undue
emphasis can be placed on the first two years of profitability, when growth profiles
are often hardest to predict. In reality, basing a European biotechnology company’s
value on its two-year projected earnings does not tend to happen too often, simply
because most European cash-burning biotechnology companies are more than two
years from profitability.
Applying discounted future earnings when net positive earnings are some years into
the future is even more fraught with problems. As discussed earlier, applying the
typically used discount rate of 25-40% is not an effective method of accounting for
clinical-stage product-development risk – particularly given that there is little
consistency in the discount rate used. This is better handled by explicitly risk adjusting
cash flows using historic product attrition rates and then applying a discount rate built
up from a risk-free rate plus a risk premium that reflects the sector’s beta.
While some analysts dismiss the product NPV build-up method as a veneer of pseudo-
scientific justification for arriving at a company’s value breakdown, it does at least have
the advantage that, having treated them all equally, each product’s relative contribution
to the company’s value is obvious to the investor. The discounted future earnings
method does not show this value breakdown nearly as clearly. Furthermore, an
argument against using probability-weighted discount rates, which is sometimes touted
by practitioners of the discounted future earnings method, is that, in reality, drug
development is a binary event: drugs either fail in clinical trials (in which case they are
worth nothing, or less than nothing if trial costs are factored in) or they reach the market
(in which case they are worth their full estimated NPV). They are never, the sceptics
contend, worth ‘30%’ of their estimated NPV. This argument is fallacious, however. The
point is that, until the outcome of the binary event occurs and is therefore known with
certainty, neither of the binary values (corresponding to success or failure) can
reasonably be assigned. Discounted future earnings do not obviate this point and neither
does using a high discount rate: if the drug fails, the company will not record the
projected earnings stream. If it does not, then there is no justification for using a 40%
discount rate. (As a footnote, the full range of actual NPV outcomes can be assessed
using Monte Carlo analysis, which we discuss later in this note. However, we shall see
that the average of these outcomes still turns out to be the risk-adjusted NPV.)
Finally, while any valuation methodology can be ‘fine-tuned’ to arrive at a desired result
that has been decided upon a priori, the discounted future earnings approach is
particularly well suited to this sort of financial alchemy. Nonetheless, it is a popular
method with specialist US institutions which, to be fair, operate in a more sophisticated
biotech market than the European one.
See back of report for important disclosures and disclaimer
19
Biotech valuation December 2004
Other methods: Comparables
By and large, without sales or earnings, it is hard to apply comparables to cash-
burning biotech companies, although they can serve as a useful common-sense check
to results derived from NPV analysis. Most common is the comparable technology
value, which examines companies by comparison with other publicly quoted
companies with notionally similar technologies, and stripping out the differences in
value attributable to cash reserves. For example, oncology companies with a set of
clinical products can compare their technology values with those of other oncology
companies with similar product profiles. Clearly, there will inevitably be numerous
differences that could account in large part for any difference in value, but the
technique can help identify outliers. However, where comparisons cannot be fairly
made vis-à-vis the stage of product development and peak sales, the use of
comparables becomes a meaningless analysis.
Comparable technology value is most commonly used for companies with partnered
enabling technologies but no clinical products of their own. It is sometimes argued that
this is a more appropriate valuation technique for platform companies than DCF
modelling. Proponents contend that risk adjusting R&D or pre-clinical-stage projects
using traditional risk-adjusted DCF modelling (if indeed a suitable attrition figure can be
found) is unfairly punitive and does not reflect the ‘true’ value of the drugs. In addition,
since many institutions do not value drugs before they reach the clinic, companies
without a clinical-stage pipeline tend not to fall within research analysts’ radars and
hence attract investor interest.
The first problem is that virtually no two technologies are similar, which undermines the
basis for establishing their values by comparison which each other. Comparable
companies are often drawn from a pool of US quoted companies, no matter how
peripheral these companies’ technologies or even business models may be to the
platform company being valued. This is done for the simple reason that US companies
tend to command much higher market valuations. Furthermore, the harsh reality, as
exemplified by the conversions of several platform companies over the last few years
to product focused ones, is that the value to be gained from platform companies is not
only hard to quantify but is also often fairly modest.
Finally, some analysts use such metrics such as cash as a proportion of market capital
to value companies. We would argue that ratios such as these can be useful when
used in conjunction with supporting information, to common-sense check a valuation or
indicate if a firm is potentially overvalued or undervalued. However, they are not
generally meaningful tools for establishing value per se.
Other methods: Options and real options
A somewhat more esoteric method of valuation that has found some favour in recent
years, particularly with regard to early-stage companies, is that of options valuation.
The term ‘real options’ is derived from the concept of applying financial options theory
to an underlying ‘real’ asset. The basis for this method is that managerial flexibility in
the face of uncertainty has value. For example, management could increase spending
on marketing for a drug when a competitor receives regulatory approval, or could
conduct further early-stage studies on a drug that might increase its chances of
success in clinical trials. Alternatively, a company may have the option to abandon a
project at some point during clinical trials, which would amount to a put option on the
remaining cash flows associated with that project. In theory, traditional DCF methods
ignore the value of that option and hence undervalue the project.
See back of report for important disclosures and disclaimer
20
Biotech valuation December 2004
The problem is that an options valuation requires the use of comparators to determine
a proxy for the asset’s volatility, which in this case will be derived from publicly traded
shares of companies with similar technology. This returns to the problem of there being
a dearth of suitable comparators. Another problem is that two common techniques
used to value real options, namely the binomial and Black-Scholes methods, share the
same conceptual flaw in that they assume the underlying value of the assets can
fluctuate along a ‘random walk’ a large number of times in between decision points. In
reality, the value of pharmaceutical R&D projects does not fluctuate substantially
between decision points and nor do they increase in value following clinical failure.
Nevertheless, the idea understandably has some popularity among platform
companies, which are commonly valued quite modestly by traditional DCF methods,
but whose values can be increased by using an options-based methodology, whereby
the options are considered call options that open up future opportunities for growth.
Indeed, one illustrious advocate of the real options valuation approach to R&D is
Merck & Co, which incorporated an option pricing approach in its portfolio investment
appraisal process in the early 1990s. Again, however, it is necessary to apply some
measure to calculate the volatility of the project, and the difficulties in obtaining a
meaningful one for biotech stocks will most likely obviate any benefit to be gained from
using the technique.
Other methods: Monte Carlo analysis
An oft-cited criticism of rNPV and other NPV methodologies is that the value
generated, be it for a product or a whole company, never actually reflects the ‘true’
value of the asset. If, by use of a hypothetical crystal ball, it were known beforehand
that a product in development would ultimately fail, the company’s value is likely to be
much lower (and the product’s NPV in isolation would be negative), whereas if it were
know beforehand that it would go on to succeed, it would be much higher. The rNPV
value is simply a weighted average of these outcomes that are, in reality, unknown.
One technique that somewhat circumvents this problem is Monte Carlo analysis. This
is a relatively straightforward process whereby a computer programme (such as a
simple Excel macro) will run a series of iterations that simulates how a product(s)
performs in trials. For example, rather than assigning a 43% chance of the product
progressing from Phase II to Phase III trials, 43% of the time the simulation will
progress the drug and 57% of the time it will model failure. This operation can be
performed for a number of different assumptions, including market-related ones: for
example, a product’s most likely market penetration might be 20%, with 15% seen as a
worst-case scenario and 25% a best-case scenario, and a normal probability
distribution curve in between. The Monte Carlo simulation will randomly choose
different values within these (and all other specified) parameters for each iteration,
repeating the process hundreds of times. A profile of the various possible actual value
outcomes can thus be built up.
While this technique is useful for assessing the spread of actual NPV outcomes for a
company depending on the various outcomes of its clinical programmes, it is not much
use for coming up with a single value for the company in the present time, over and
above that which is provided by traditional NPV analysis. Of course, it can serve to
exemplify the risks in, for example, a company with one high-risk clinical programme
with considerable peak sales if successful. This risk-concentrated company might yield
a healthy rNPV, but is otherwise akin to buying a lottery ticket. Having said that, most
biotech investors are likely to appreciate this without the aid of Monte Carlo analysis.
See back of report for important disclosures and disclaimer
21
Biotech valuation December 2004
Fig 11 Monte Carlo simulation
0
50
100
150
200
250
300
350
-75
2
5
1
25
2
2
5
3
2
5
425
525
6
25
7
25
8
2
5
9
2
5
1
,02
5
1
,12
5
1,225
1,325
1,4
2
5
NPV (US$m)
Frequency of outcome
Note: Assumes a company with one Phase III product and forecast peak sales of US$500m.
Source: ING estimates
_
Sales forecasting
The basis for all self-explicative valuations is a product sales forecast. Not only does
the forecast provide the basis for estimating revenues, but the forecasting process
itself drives an understanding of the value of the company’s core tangible assets – its
future drugs.
In its most advanced forms, sales forecasting is a marketing discipline rather than a
financial one, and a detailed treatment of the subject is beyond the scope of this report.
Major pharmaceutical companies often employ a suite of sophisticated techniques in
order to forecast future product sales, although at a fundamental level the process
always relies on gauging the likely level of demand for the product. This, in turn, will
depend on:

Prevalence, incidence and demographic/age segmentation of the target patient
population;

Prevalence and incidence of any co-morbidities (ie, related target markets), plus
potential off-label use;

Diagnosis and treatment rates;

Nature of the target diseases (eg, chronic, acute, life threatening, etc);

Degree and nature of the product’s clinical efficacy, including the efficacy
advantage conferred over existing standard-of-care treatments;

Treatment regimen, including dosing frequency, route of administration, side-effect
profile, etc (from which a patient compliance rate can be estimated);

Marketed and development-stage competition; and

Likely product pricing and reimbursement status.
Clearly, some of these factors can be ascertained with a reasonable degree of comfort
from publicly available data sources (such as disease prevalence), while others (such
as clinical efficacy for early-stage drugs) cannot be known with any certainty. In this
latter case, assumptions must be made according to the best available data and on the
See back of report for important disclosures and disclaimer
22
Biotech valuation December 2004
basis that the drug will perform according to reasonable expectations, and modified
subsequently as more clinical data comes to light.
Using all of these assumptions as a basis, the next step in the forecasting process is to
construct an epidemiological model, ideally using age-segmented prevalence and
incidence data, that will assess the level of market penetration within the treatable
target populations, and also the duration, frequency of treatment and compliance rate
per patient.
How do you estimate market share?
Estimating the degree of market penetration that a drug is likely to achieve is
one of the hardest parts of the forecasting process. In some cases, the market
share that drugs directed at established targets are likely to take from the first-
in-class market leader follows a fairly well documented inverse power rule, so
that analysis of sales data from a bank of historical surrogate drugs can
sometimes provide a useful predictor of likely sales of fast followers, or mildly
differentiated ‘me-too’ drugs. By extension, it is not unreasonable for the time-
constrained investor to estimate the likely market share of a follow-on drug by
taking a percentage of the market leader’s sales and adjusting it for other
factors. However, this is a far from infallible process and it is crucial to recognise
this (for example, Glaxo Wellcome’s anti-ulcerant Zantac outsold SmithKline
Beecham’s Tagamet, despite reaching the market place some time later and
offering questionable clinical superiority). Furthermore, this form of ‘top down’,
or sales-based, forecasting is generally ineffective for highly novel drugs that
are likely to significantly expand, or even create, their market (such as Pfizer’s
Viagra). This is particularly pertinent to biotech companies, which are often
developing NCEs directed at novel targets. Even if the target is an established
and validated one, it can be difficult to distinguish a low value, me-too follow on
from a truly superior therapy. For example, Pfizer’s Lipitor, the third statin to
market, came to be regarded as a ‘next generation’ statin and became the
world’s biggest selling drug.
The most reliable way to estimate market share is to perform some form of
quantitative market research, and hence ascertain the likely demand for the
product based on its attributes such as efficacy, dosing, side-effect profile and
so on. The results should then be considered in light of the competitive analysis,
ie, a drug with good clinical results compared with standard of care could clearly
be impacted significantly by a development-stage product with better results.
This market research can range from the fairly basic, such as interviewing a
handful of clinicians, up to conducting several hundred physician interviews to
power what is known as a conjoint model. This is a market research technique
that models how a product’s attributes, which are termed its ‘part worths’, are
likely to be valued by the market. When these results are fed into buyer choice
simulations, or compared with historical data from existing marketed drugs, a
reasonable estimate of market share can be made. Clearly, this particular
technique is expensive, time-consuming and not particularly suitable for the
investor looking to value a whole portfolio of products within a company.
However, a rigorous (and, ideally, quantitative) survey of physician opinion –
both opinion leader and regular clinician – is a prerequisite for properly
understanding a drug’s prospects based on its known attributes.
See back of report for important disclosures and disclaimer
23
Biotech valuation December 2004
It is important to stress that a professional market forecaster will be thorough in the
construction of the patient waterfall model and will not simply determine the number of
diseased patients and then estimate peak penetration (note that the term ‘market
penetration’ is often incorrectly used in this context – strictly speaking, market share
refers to the share of US dollar sales a product can obtain in an established market).
Some age groups may be contra-indicated pharmaceutical treatment, some product
types may not be prescribed continuously (this is know as the continuation rate), and
so on. While this degree of detail is not practical for most investors, it demonstrates the
point that many of these assumptions are frequently ignored and implicitly overstated,
resulting in analysts’ sales forecasts often being unrealistically high.
We assign a standard, roughly sigmoidal growth curve for each product up to a peak
level of sales or patients treated. Some researchers use a greater number of
classifications to describe the lifecycle of a product (for example, short, medium or
long), which differ in the rates of growth and decline of the drug. At the most
sophisticated level, pharma forecasters model the uptake of a drug among a
population of prescribers using product diffusion models. Despite this, we find that it is
not normally practical or possible to attempt to do this and, in the absence of modifying
information, a standard growth curve usually suffices.
Finally, some attention must be paid to the likely treatment price and reimbursement
status of the product. With increasingly cost-conscious healthcare reimbursement
systems, including in the US, assuming premium annual treatment costs for a product
offering modest improvements over standard-of-care treatment may be unwise.
Fig 12 Gauging product risk

• Not GMP
• Scale up not achieved
• Capacity constraints
Other risks

• Low cost scale-up
achievable
• Manufacturing cGMP

• Ethical or political issues concerned with
product or technology
• New therapeutic approach
Regulatory risk

• FDA/EMEA approved trial design
• Political pressure to expedite approvals

• Unknown mode of action
• Un-validated by competition
Molecular risk

• Solid pre-clinical studies
• Reformulation of approved active
• Mechanism of action validated by approved
product
Product risk

• Clinical trials in very limited
numbers of patients
• Physician-led, open-label trial as
proof of concept
• Equivocal/unexplained trial results
• Clinical development plan is
significantly different from similar
approved products
Clinical risk

• Statistically powered
studies
• Double-blind,
randomised, controlled
• Trials conducted to GCP
• Trials designed in
conjunction with
regulators
Source: ING
_
See back of report for important disclosures and disclaimer
24
Biotech valuation December 2004
How it works in reality: Target prices
We have constructed a standardised NPV for a hypothetical company with one product
with forecast peak sales of US$100m.
Fig 13 Standardised NPV model for Phase II product (launched four years from year 0)
Launch Year 2 Year 3 Year 4 Year 5
Sales (US$m) 20 41 61 82 100
Growth (%) 105 49 34 22
Royalty (%) 15 15 15 15 15
Company share (US$m) 3.0 6.2 9.2 12.3 15.0
Overhead (%) 30 30 30 30 30
Net operating profit (US$m) 2.1 4.3 6.4 8.6 10.5
Tax (%) 0 0 30 30 30
After-tax profit (US$m) 2.1 4.3 4.5 6.0 7.4
Risk-adjusted profit (US$m) 0.6 1.3 1.3 1.8 2.2
Terminal value (20x multiple) (US$m) 44.1
NPV (cash flows and terminal value discounted back to present day) (US$m) 16.89
Source: ING estimates
_
In practice, this NPV of US$16m will not be the same value as the company’s actual
market capitalisation. But is it the value that the company should trade at? We noted
earlier that fair values derived from NPV analysis (or any other valuation methodology)
almost never match actual share prices, which reflects a dynamic equilibrium of market
supply and demand. Part of this is due to error endogenous to the NPV – any single
analyst’s valuation is ultimately only an opinion of value, which will invariably differ
from those of other analysts, even if not by much. Part of it is reflective of sentiment,
both to equities in general and the sector specifically. In 2000, most biotech companies
traded at relatively narrow discounts to their fair values (10-20%), while for the last few
years they have traded at much wider discounts, in the range of 30-40%. It can be
seen that this discount has varied over the past few years, largely reflecting changes in
sentiment toward the sector in response to wider macroeconomic conditions.
Fig 14 Discount to NPV – historical
-58%
-44%
-52%
-37%
-36%
-60%
-55%
-50%
-45%
-40%
-35%
-30%
12/1/01 12/1/02 12/1/03 12/1/04 12/1/05
Discount to NPV
Source: ING
_
Finally, part of it will reflect the immutable truth that there are numerous qualitative
factors relating to a company that cannot readily be measured in any model. Quality of
management is one obvious one: a company that has a history of over-promising and
underperforming, burning shareholders’ cash with little apparent benefit and generally
alienating its shareholder base is clearly going to warrant a wider discount by the
See back of report for important disclosures and disclaimer
25
Biotech valuation December 2004
market. Equally, companies that have conducted a recent fundraising and which will
therefore have temporarily exhausted buying demand, or companies suffering
disappointing newsflow, will also trade at wide discounts.
Nevertheless, given the fact that nearly all cash-burning companies trade at discounts
to their NPVs, the market currently does not value these companies according to
exactly the precise same logic. Why is this? There are probably several reasons,
including:

Cash-burning companies’ cash reserves are discounted from their current levels to
reflect committed burn;

The market views the development risks relating to most biotech companies’
products as being greater than ‘traditional’ pharma attrition rates would suggest;
and

In an environment where equities – even in defensive sectors – are not in favour,
risk stocks such as biotechs warrant an additional discount.
For instance, if we were to apply the average sector discount of 30-40% to the NPVs of
Phase III products, we would see an implied Phase III probability of reaching the
market of 45%, which is not dissimilar to the actual levels of attrition seen in the Euro
biotech universe. While we would not claim that this is necessarily anything other than
coincidence, it does suggest that the market may already take into account an above-
normal level of biotech product development risk for European biotechs.
A full set of implied attrition rates is shown in Figure 15 assuming a 40% discount rate.
Fig 15 Clinical trial success rates
Phases Phase to phase (%) Phase to market (%)
Pre-clinical 50 3
Phase I 33 6
Phase II 43 18
Phase III 78 42
Registration 90 90
Market 100
Source: ING estimates
_
We present here a simplified method for arriving at feasible target trading prices based
on a variety of routine parameters. The starting point is to discount each development-
stage product and cash reserves by an average sector discount to NPV (ie, by around
37% based on current values). Note that we do not discount products that are already
marketed and generating revenues, on the basis that the risk premium ascribed by the
market to biotech companies will no longer be warranted for these products.
Figure 16 shows the discounts to NPV of the cash-burning companies in our European
biotech universe at the time of publication.
See back of report for important disclosures and disclaimer
26
Biotech valuation December 2004
Fig 16 Discount to NPV
-80% -60% -40% -20% 0% 20% 40%
Crucell
GW Pharmaceuticals
Alizyme
NeuroSearch
GPC Biotech
CaT
Vernalis
Basilea
Genmab
ARK Therapeutics
Zeltia
Average
Phytopharm
NeuTec Pharma
Pharmagene
SR Pharma
Pharmexa
CeNes
Medigene
NicOx
Antisoma
Xenova
Discount to NPV
Source: ING
_
Interestingly, when this discount is applied to the cash balance, it approximates to one
year’s burn rate for the ‘average’ biotech (the average cash life of our cash-burning
biotech universe is 33 months, so one year’s burn represents a reduction of 36% of the
average cash balance). This information can also be used to assess how the average
European biotech’s NPV should evolve year by year in the absence of newsflow or
fund raisings. It can be seen that the average European biotechnology company’s NPV
comprises 20% cash (which, as noted before, amounts to 33 months burn) and 80%
technology value. Hence, one year from now, its product NPV will have increased by
13%, the standard discount rate, as the pipeline matures by one year and, in the
absence of any fundraising, its cash will have decreased by the equivalent of one
year’s burn, or 36%. The net effect on the company’s NPV is shown in Figure 17
assuming the average European biotech has a market cap of €100m.
Fig 17 Average European biotech company’s one-year NPV evolution (€m)
Now One year
Products/technology 80 90
Cash 20 12
NPV 100 102
Source: ING
_
In effect, the NPV stays more or less the same. It is important to stress that these are
clearly no hard-and-fast figures, and this serve merely as a starting point for further
analysis. A company that is very cautious with its reserves may warrant a smaller
discount than one that is not. In addition, since average discount-to-NPV figures are
being used to arrive at trading prices, some companies will inevitably have a
See back of report for important disclosures and disclaimer
27
Biotech valuation December 2004
theoretical trading price above their current market price and some below it. The key is
that it provides a useful starting point from which, when examined in conjunction with
qualitative factors, companies that are potentially overvalued or undervalued with
respect to their peer group can be identified.
Fig 18 Cash survival index
0 10 20 30 40 50 60 70 80 90 100
Vernalis
Xenova
Phytopharm
SR Pharma
Innogenetics
GW Pharmaceuticals
Alizyme
XTL Biopharmaceuticals
Pharmagene
Pharmexa
Oxford Biomedica
Zeltia
ARK Therapeutics
Evotec OAI
Cytos
NicOx
Antisoma
Average
CaT
CeNeS
Medigene
Genmab
Crucell
GPC Biotech
Basilea
NeuroSearch
Morphosys
Remaining cash (months')
Source: Company data
_
Conclusions
The method for arriving at trading ranges presented above is clearly based on
approximations and can only serve as a rough-and-ready guide. However, it is
frequently seen to yield relatively accurate results in practice. Examples that highlight
the utility of our preferred approach can be seen among recent European IPOs, where
we have seen a trend toward aggressive pricing. Two of these, Ark Therapeutics and
Basilea, both suffered price declines shortly after flotation, retracing toward our
predicted price range. Ark listed at £1.33 per share, against our NPV valuation (arrived
at using the standard methodology with peak sales according to the lead bank’s
estimates) of £1.26. Applying our methodology for discounting to arrive at a fair trading
price yields a value of 84p, not far from the 86p per share price at time of publication.
Similarly, Basilea listed at SFr98, against our standard NPV of SFr104 and trading
price of SFr68. The current price is SFr78, although this has risen on the back of
positive newsflow from a low of around SFr62 not long after flotation.
See back of report for important disclosures and disclaimer
28
Biotech valuation December 2004
Sensitivity analysis, scenario
analysis and NPV evolution
A key requirement of NPV analysis is that the sensitivities of different assumptions be
understood. Broadly speaking, sensitivity analysis considers the changes in valuations
due to flexing individual assumptions in isolation. By comparison, scenario analysis
looks at the linkages between different assumptions and assesses the impact of
various combinations of assumptions. As an example of the former, many analysts
look at variation in NPV using different discount rates. An equally useful sensitivity
analysis could look at variation in NPV with peak sales of key products or risk-
adjustment rates. There is, nevertheless, quite a high degree of subjectivity in doing
this, as it is necessary to try and maximise the accuracy of the starting assumption and
keep the sensitivity boundaries fairly narrow, or else the valuation range will be
unhelpfully large and leave the user with little real idea of what the company is worth.
Scenario analysis
Scenario analysis is perhaps more useful in practice, as the outcomes of events can
often be modelled quite realistically, and their impact on the company’s value
assessed. In certain cases, the likely outcomes have to be estimated (for example, the
possible different results of a patent dispute), while in others they are binary events,
such as product success of failure in a clinical trial. A number of different scenarios can
be fairly readily modelled

The outcome of a product failure or success.

The entry of a new drug into the clinic.

The entry of existing products into new indications.

Additional sales potential.

The outcome of a fundraising.

The outcome of a royalty dispute.
Example 1
An example of the first case is that of NeuroSearch. In September 2004, we assessed
how NeuroSearch’s NPV would evolve depending on the clinical outcomes of two of its
drugs. One was its ADHD drug NS2359 (which later went on to fail) and the second its
combined Alzheimer’s/Parkinson’s drug, NS2330. The following NPV outcomes could
be shown depending on the clinical results from these drugs.
See back of report for important disclosures and disclaimer
29
Biotech valuation December 2004
Scenario 1: NS2359 fails in ADHD Phase II trials
Fig 19 Evolution of NPV (DKr) if NS2359 fails in ADHD
226 Probability = 57%
94
203
292
401
NS2330 fails in
Alzheimer’s and
Parkinson’s
NS2330
successful in
Parkinson’s but
not Alzheimer’s
NS2330
successful in
Alzheimer’s but
not Parkinson’s
NS2330
successful in
both Alzheimer’s
and Parkinson’s
P = 32% P = 25% P = 25% P = 18%
NPV below current fair value
NPV above current fair value
Source: ING
_
Scenario 2: NS2359 is successful in Phase II trials
Fig 20 Evolution of NPV (DKr) if NS2359 results are positive in ADHD
486 Probability = 43%
354
463
552
661
NS2330 fails in
Alzheimer’s and
Parkinson’s
NS2330
successful in
Parkinson’s but
not Alzheimer’s
NS2330
successful in
Alzheimer’s but
not Parkinson’s
NS2330
successful in
both Alzheimer’s
and Parkinson’s
P = 32% P = 25% P = 25% P = 18%
NPV below current fair value
NPV above current fair value
Source: ING
_
Example 2
A second example looked at the outcome of Cambridge Antibody Technology’s dispute
with Abbott Laboratories regarding the royalty stream due on CAT’s rheumatoid
arthritis drug, Humira. A reasonable upside and downside case could be considered,
and the likely results on NPV modelled.
We have conducted a scenario analysis of the potential impact of both a positive and
negative outcome of the royalty dispute on our valuation.
Should CAT win the case, we estimate its net royalty will increase threefold, from 1%
to 3%.
See back of report for important disclosures and disclaimer
30
Biotech valuation December 2004

CAT succeeds in royalty dispute. Our assumption of net royalty on Humira sales
would increase to 3.0% contributing 476p to our NPV valuation of 956p.

CAT fails in royalty dispute. Our assumption of net royalty on Humira sales would
remain unchanged at 1%, contributing 159p to our NPV valuation of 619p.

CAT fails in royalty dispute, but only Cabilly is offset. Our assumption of net
royalty on Humira sales would increase to between 1.33% and 2.0%, contributing
between 211p and 317p to our NPV valuation of between 675p and 787p.
Fig 21 Humira court case scenario analysis
Scenario NPV Humira (p) Total NPV fair value (p)
NPV/share if CAT loses court case 159 619
NPV/share if CAT wins court case 476 956
NPV/share if CAT loses case but offset limited to Cabilly 211-317 675-787
Source: ING
_/
Fig 22 Scenario analysis
159p
159p
619p
458p
-23%
619p
956p
788p
+31%
211p - 317p
675p - 787p
512p - 623p
+8% to -14%
Humira NPV
Total CAT NPV
Trading range
35% discount*
Downside
Upside/Downside
Upside
CAT fails in court case -
minimum royalty of 1%
CAT wins court case - net
royalty increases to 3%
CAT loses but only Cabilly
offset - royalty increases to
between 1.3 and 2.0%
Note: We apply an average sector discount to NPV of 30-40%; however, note that we do not discount Humira royalties as the product is approved. Upside and
downside are calculated based on CAT’s price of 600p on 29 September 2004.
Source: ING
_
Conclusion
Another reasonably predictable event that can be forecast is company fund raisings.
When a company comes within 18 months of its cash reserves, based on current and
forecast burn rates, it is likely that it will have to undertake a cash call for at least two
years’ worth of further funding. In this instance, we can adjust the company’s NPV by
assuming it raises two years’ worth of cash at a 15-30% discount to its share price,
and model how the company’s NPV changes as a result.
The value of this type of analysis is chiefly that the outcomes of major value-changing
events can be assessed well in advance, so that when the event occurs and the share
price adjusts, it can be gauge to what degree the market has reacted ‘correctly’.
See back of report for important disclosures and disclaimer
31
Biotech valuation December 2004
Profitable companies
Profitable companies, a relatively rare commodity in European biotechnology (although
common in the US), are companies with net profit-generating products. These
companies are differentiated from the rest of the biotechnology sector because they
have a track record of profitability but, equally, have forecasted growth rates that are
well above those seen in the pharmaceutical industry. Generally, there are two main
classes of profitable company.
I. Specialty pharma companies (eg, Shire, Lundbeck, Warner Chilcott)
II. High-growth biotechs, of which there are, in turn, two main types.
• Long-established biotechs, such as Serono.
• Biotechs that have crossed the threshold to profitability on the back of one
major product approval (eg, Actelion).
Overview of techniques
Valuing profitable companies is arguably a much easier task than valuing cash-burning
biotechs, since the former have a history of tangible earnings and can therefore be
assessed using commonly accepted and traditional valuation methodologies. These
generally include earnings-based techniques such as PER, EV/EBITDA, EV/EBIT and
PEG ratios.
The first obvious point to note regarding each of these techniques is that they are
relative, rather than absolute, measures of value. This tends to be much less of a
problem with profitable companies than unprofitable ones, since the market is
comfortable with the comparisons being made where earnings are visible.
Although we look at each of these methods in valuing specialty pharma/FIPCOs, our
preferred ratio is the PEG ratio, because it allows us to make direct growth
comparisons, which should be a key driver of value. A PEG is equal to the prospective
PE multiple divided by the average forecast growth in earnings. A PEG can be used
most successfully as a measure of relative value when comparing growth stocks within
a sector. The following tables set out the PEG ratios and other earnings multiples for
both specialty pharma and profitable biotech companies.
A limitation of PEG is that, in order to come up with a meaningful earnings growth rate,
compound annual growth must be measured over a sufficient time span. We would
typically examine a five-year forward CAGR for this purpose. This leaves open the
risks inherent in forecasting, however, the valuation’s sensitivity to forecast error tends
to be far lower than it is in the case of DCF valuations of biotechs.
When we plot PE ratios against five-year CAGR in earnings, we can establish a line of
best fit for the constituent companies. The gradient of the line then represents the PEG
ratio for those companies (note that this will generally be different to the mean). From
this, we can make a reasonable inference that those companies that fall below the line
are, in the absence of other factors, typically undervalued by the market, and those
above the line are overvalued.
See back of report for important disclosures and disclaimer
32
Biotech valuation December 2004
Fig 23 Valuations of global specialty pharmaceutical and profitable biotechnology companies
Company Share price
(lc)
Market cap
(US$bn)
LT CAGR
(%)
2005F PE
R
(x)
2005F EPS
(lc)
2005F PEG
(x)
Allergan (US$) 78.94 10.60 19 24.4 3.23 1.3
Forest Laboratories (US$) 42.61 15.78 19 16.0 2.66 0.8
Ivax Corporation (US$) 14.82 3.72 24 17.4 0.85 0.7
King Pharmaceuticals (US$) 11.54 2.79 9 11.7 0.99 1.3
Lundbeck (DKr) 105.00 4.29 14 15.0 7.02 1.1
Shire Pharmaceuticals (£) 5.35 4.79 7 14.1 0.38 2.1
Teva Pharmaceutical Ind (US$) 26.99 16.85 21 16.6 1.63 0.8
Watson (US$) 28.65 3.14 10 14.4 1.99 1.4
Average 15 1.2
Amgen (US$) 59.84 76.00 20 24.93 2.40 1.22
Biogen IDEC (US$) 58.31 19.45 20 34.06 1.712 1.69
Cephalon (US$) 47.22 2.72 25 16.32 2.89 0.65
Chiron (US$) 31.37 5.86 15 16.36 1.918 1.09
Genentech (US$) 49.07 51.50 28 44.45 1.104 1.60
Genzyme (US$) 55.17 12.66 17 26.65 2.07 1.60
Gilead (US$) 35.16 15.22 24 30.39 1.16 1.28
Medimmune (US$) 26.23 6.52 25 61.72 0.425 2.50
Serono (SFr) 754 9.98 13 16.04 47.00 1.23
Average 20.7 30.10 1.43
Source: Company data, ING estimates
_
Fig 24 Specialty pharmaceutical valuations
Fig 25 Profitable biotech valuations
Teva
Allergan
Forest
IVAX
Watson
Lundbeck
Shire
King
9
11
13
15
17
19
21
23
25
27
5% 10% 15% 20% 25%
Serono
Medimmune
Cephalon
Gilead
Amgen
Chiron
Biogen IDEC
Genzyme
10
20
30
40
50
60
70
12% 14% 16% 18% 20% 22% 24% 26%
Source: ING estimates Source: ING estimates
Pharma stocks have traditionally traded around a PE ratio of 20-25x with profitable
biotechs trading at 1.5-2.0x the pharma multiple. This resulted in large biotech
having a lower PEG than pharma given pharma’s low double-digit EPS growth and
large biotech’s 20%-plus growth. Pharma is currently trading on a PEG of 1.35x
whereas large biotech is on a 1.43x PEG. Given that large biotechs have
historically achieved PERs of 30-50x and have seen PEGs in the 1.5-2.0x range,
they are currently trading at the bottom of that range.
Specialty pharma stocks have traditionally traded at a discount to the
pharmaceutical sector on a PEG basis, with average multiples of 1.2-1.3x,
reflecting the slightly riskier earnings profile of those stocks (eg, over-reliance on
one product). The current PEG for these stocks is also at the bottom end of the
range at 1.2x
.
See back of report for important disclosures and disclaimer
33
Biotech valuation December 2004
Emerging profitable companies
‘Emerging profitable’ companies are those companies that are on the cusp of
profitability; that is, they can reasonably be expected to generate solid net earnings
within three years. Examples would include companies such as SkyePharma or Elan.
Owing to their somewhat hazy position in the valuation spectrum, different valuation
metrics have varying degrees of applicability to these companies.
Traditional metrics, such as PE ratios, are generally not applicable to companies that
lack earnings, for many of the reasons outlined earlier. However, companies close to
profitability or breakeven, with products on the market, must be valued using a variety
of techniques including:

DCF or risk-adjusted DCF analysis;

Discounted earnings; and

Comparable EV sales analysis can be useful in some circumstances, although we
use this less widely.
Example 1
As an example, we include our valuation of Elan published on 09 November 2004 –
please note this is for descriptive purposes only and does not reflect our current
valuation or earnings estimates. In this valuation, we utilised a mix of DCF, sum-of-the-
parts NPV analysis and discounted earnings.
DCF analysis
Our key valuation technique for Elan is based on a DCF. We believe this valuation
technique gives the correct balance of value to the existing products and the potential
for long-term value within Elan. Our model is based on our current estimates for growth
from current products to 2010F. Outside this forecast period, we fade growth from 11%
in 2010F to 3.5% in perpetuity from 2022F. We utilise a calculated discount rate of
9.3% and assume a 3.5% growth rate into perpetuity.
Assumptions

Sales of Antegren reach US$3.1bn in 2010F in multiple sclerosis alone. Antegren is
launched in early 2005F.

Sales of Antegren reach US$500m in Crohn’s and is launched in 2006F.

Our DCF valuation using these assumptions suggests a fair value of US$28.8
(previously US$25.0).
See back of report for important disclosures and disclaimer
34
Biotech valuation December 2004
Fig 26 Free cash flow (US$m)
2004F 2005F 2006F 2007F 2008F 2009F
Turnover 481.9 736.9 1,260.4 1,900.4 2,508.5 2,883.8
-Turnover growth (%) 53 71 51 32 1
5
EBIT (308.4) (56.8) 75.2 448.7 732.3 889.9
Operating margin (%) -64 -
8
6
24 29 31
Tax-effected EBIT (308.4) (57.3) 76.0 426.3 695.7 845.4
+ Depreciation 123.0 134.3 121.3 112.6 108.9 70.9
- Capital expenditure (38.6) (51.6) (75.6) (95.0) (125.4) (144.2)
+ Decreases/- incl in wk cap (91.8) 2.2 51.2 8.5 (16.7) (17.2)
Free cash flow (315.7) 27.6 172.8 452.3 662.4 755.0
Source: ING estimates
_
Fig 27 DCF analysis (US$m)
Perpetuity growth method
2.5% 3.0% 3.5% 4.0%
Present value of cash flows 7,065.4 7,065.4 7,065.4 7,065.4
Present value of terminal value 5,143.1 5,578.6 6,089.2 6,696.2
Total enterprise value 12,208.5 12,644.0 13,154.6 13,761.7
- (Net debt)/+ net cash* (87.5) (87.5) (87.5) (87.5)
Equity value 12,121.0 12,556.5 13,067.1 13,674.2
Price per share as per DCF analysis (US$) 26.73 27.69
28.81 30.15
*Our calculations utilise net debt excluding the US$460m convertible. As this bond is currently in the money, we
utilise a diluted no. of shares in our valuation of 453.5m.
Source: ING estimates
Sum-of-the-parts valuation
Our sum-of-the-parts NPV gives a value of US$17.35 per share on Elan’s current
pipeline. At current levels, the shares are therefore trading at a premium of 67% to our
sum-of-the-parts NPV.
Fig 28 Sum-of-the-parts valuation (US$m)
Valuation Diluted
Senior notes (7.25%) 650.0 650.0
Convertible notes (6.5%) 460.0
Fixed payments 1.9 1.9
EPIL III 390.0 390.0
LYONs 0.9 0.9
Lease obligations 78.9 78.9
Operating lease obligations 141.7 141.7
Total 1,723.4 1,263.4
Cash at 30.06.04 680.3 680.3
Investments at 30% liquidity discount 275.0 275.0
Net debt/commitments (768.1) (308.1)
Current products and DD business 1,415 1,415.4
Prialt valuation 370 370.0
Antegren valuation 6,391.0 6,391.0
Total valuation 7,407.9 7,868
Shares – fully diluted (m) 391.00 447.0
Elan valuation (US$) 18.94 17.60
Source: ING estimates
_
See back of report for important disclosures and disclaimer
35
Biotech valuation December 2004
Discounted earnings
In addition to our DCF analysis, we have conducted a discounted earnings valuation
based on our 2008 earnings estimates discounted back three years to give a 12-month
price target range. Given that earnings will be growing at a 2007-10F CAGR of 38%,
we feel comfortable that Elan’s shares will be trading at a 35x multiple or above, in line
with its US biotech peers.
Fig 29 Discounted earnings (US$)
10% 15% 20% 25%
30.0x 31.8 27.8 24.5 21.7
35.0x 37.1 32.5 28.6 25.3
40.0x 42.4 37.1 32.7 28.9
45.0x 47.7 41.8 36.8 32.5
Source: ING
See back of report for important disclosures and disclaimer
36
Biotech valuation December 2004
Platform technology and tools
companies
Although we consider platform technologies and so-called ‘tools’ companies (ie, the
reagents, in vitro diagnostics and consumables companies) in the same chapter, the
obvious point needs to be made at the outset that there can be profound differences
between the two classes of company. An early-stage genomics company with low
revenues and heavy losses is clearly fundamentally different from a large consumables
company such as Qiagen. The issue that needs to be examined is whether platform
companies seek to generate value from their technologies based on a service model,
the development of a proprietary pipeline, or a combination of the two.
Platform companies
Platform companies include those companies that aim either to expedite or improve the
pharmaceutical R&D process based on the application of a proprietary technology, or to
generate novel compounds in an entirely new way. Historically, these technologies have
included systems such as bioinformatics, combinatorial chemistry, parallel screening,
genomics and pharmacogenomics, cell lines, antibody generation and so forth. Lately,
we have also seen the emergence of platforms that enable in silico lead generation,
selection, structural modification or pharmacokinetic and toxicological screening. In many
cases, the benefits that these technologies offer the pharmaceutical or biotechnology
R&D process are not in doubt, and some technologies (such as combinatorial chemistry)
eventually establish themselves as routine components of the canon of drug
development processes. The more difficult problem arises in trying to extract tangible
value from a system that will over time become a commodity service and ultimately
redundant, and in trying to quantify that value.
Strict application of the NPV methodology does not yield significant valuations for most
technologies. An argument can be presented that demonstrates how the technology might
improve pre-clinical productivity from traditional drug discovery, shown in Figure 30.
See back of report for important disclosures and disclaimer
37
Biotech valuation December 2004
Fig 30 The drug discovery process
What?
Hit to lead screening
Efficacy
Specificity
Bio-
availability
Early stage development
Late stage
development
Lead optimisation
Pre-clinical development
Lead generation
Phase I
Phase II
10,000
compounds
5,000
leads
250
optimised
leads
20 10
Compounds
Probability
of success
0.01% 0.02% 0.4% 5%
10
%
18 months 24 months30 months
Traditional
Timing
5
years
to
market
4
years
to
market
New
technology
implications
Probability
of success
0.05% 1.1% 5%
10
%
0.03%
No of
compounds
1,800 90 20 103,500
6 months 18 months 18 months
New
timing
Traditional
New
technology
12 12
6 12
5
years
to
market
4
years
to
market
Source: ING
_
This scenario demonstrates an improvement, both in attrition rates and timelines, that
could theoretically be obtained with a novel platform technology. However, a key point
here is that, despite the improvement in lead generation, clinical-stage attrition rates
have not changed. Therefore, following the logic of this exercise and incorporating
suitably adjusted pre-clinical attrition rates into the NPV model, we still, as exemplified in
Figure 31, arrive at comparatively modest valuations for the early-stage projects.
Fig 31 NPV evolution of US$500m product (US$m)
5% royalty 10% royalty 15% royalty
NPV at lead generation 0.04 0.08 0.12
NPV at lead optimisation 0.98 1.96 2.94
NPV at early pre-clinical 2.84 5.69 8.53
NPV at late pre-clinical 3.94 7.88 11.83
NPV at Phase I 7.88 15.77 23.65
NPV at Phase II 26.73 53.45 80.18
NPV at Phase III 70.47 140.94 211.41
Source: ING
_
See back of report for important disclosures and disclaimer
38
Biotech valuation December 2004
Fig 32 NPV evolution of product bearing a 10% royalty
0.04
0.98
2.84
3.94
7.88
26.73
79.63
0
10
20
30
40
50
60
70
80
90
L
ea
d
g
en
e
ration
L
ea
d
o
pt
i
m
i
s
a
ti
o
n
E
a
rly p
r
e
-c
l
i
n
i
c
a
l
L
at
e
pre
-
cl
i
ni
cal
P
h
a
s
e
I
P
h
a
s
e
II
P
h
a
s
e
II
I
Value ($m)
Source: ING
_
Fig 33 NPV evolution of product bearing a 20% royalty
0.08
1.96
5.69
7.88
15.77
53.45
159.26
0
20
40
60
80
100
120
140
160
180
L
ea
d
g
en
e
ration
L
ea
d
o
pt
i
m
i
s
a
ti
o
n
E
a
rly p
r
e
-c
l
i
n
i
c
a
l
L
at
e
pre
-
cl
i
ni
cal
P
h
a
s
e
I
P
h
a
s
e
II
P
h
a
s
e
II
I
Value ($m)
Source: ING
_
Fig 34 NPV evolution of product bearing a 30% royalty
0.12
2.94
8.53
11.83
23.65
80.18
238.89
0
50
100
150
200
250
300
L
ea
d
g
en
e
ration
L
ea
d
o
pt
i
m
i
s
a
ti
o
n
E
a
rly p
r
e
-c
l
i
n
i
c
a
l
L
at
e
pre
-
cl
i
ni
cal
P
h
a
s
e
I
P
h
a
s
e
II
P
h
a
s
e
II
I
Value ($m)
Source: ING
_
The most obvious alternative approach (if the application of real options methodology
is to be accepted as being too arcane for most investors) is to value the technology on
the basis that it is partnered on a service model to generate FTE revenue, which most
See back of report for important disclosures and disclaimer
39
Biotech valuation December 2004
companies will do whether or not they are developing proprietary products. In this
case, the revenue stream thus generated can be valued on a revenue multiple basis,
which provides some sort of comparable valuation on which to proceed. In this case,
we would derive appropriate figures using EV/sales ratios from comparable
companies. Note that the use of comparables is appropriate in this case, because a
tangible revenue stream is being valued, irrespective of the nature of the underlying
platform technology. The argument can be made that the value of the technology to
pharmaceutical partners is implicit in the revenues it generates from collaborations.
Indeed, some European companies (such as Morphosys) generate relatively healthy
and sustainable revenues in this way. An alternative approach, which uses
comparables based on the market capitalisations (less cash) of other publicly traded
technology companies, is less appropriate, since finding truly comparable technologies
is difficult.
The advantage of the EV/sales approach is that it can be used in combination with the
standardised NPV approach to value the different parts of a hybrid business. For
example, for companies such as Morphosys or Pharmagene, both products and the
service businesses can be valued and combined to derive an overall fair value for the
company.
As an example to highlight the valuation comparisons for technology platform
companies that we use, we believe that the chemistry and biology services universe
represent good case studies. As these companies are primarily fee for service, the
standard valuation methodology is EV/sales one year out. However, the multiples by
themselves do not truly reflect how the market values the individual companies.
The market also considers the quality and sustainability of the businesses and thus
future growth rates at the top line. A comparison of EV/sales to sales growth is an
equivalent PEG analysis for companies that are valued on the top line.
Fig 35 Valuations of global chemistry services companies
Company Cur
r
Share price Mkt cap EV 2004F sales EV/sales (x)
(lc) (lcm) (lcm) (lc m) 2005F 2004F 2005F
Albany Molecular Research US$ 10.92 346.8 407.3 168.1 167.1 2.42 2.44
ArQule US$ 5.41 156.3 67.5 52.4 51 1.29 1.32
Array Biopharma US$ 8.46 245.2 191.5 35 39 5.47 4.91
Biofocus £ 100 16.3 17.9 17 20.2 1.05 0.89
Cerep € 10.07 120.5 88.6 51 60 1.74 1.48
Discovery Partners Intl US$ 4.65 121.5 79.2 53.4 59 1.48 1.34
Evotec OAI € 2.93 104 176.4 79 87 2.23 2.03
Pharmacopeia Drug Discovery US$ 5.8 71.4 26.4 29.4 30 0.9 0.88
Tripos US$ 5.01 46.6 73 N/
A
N/A N/A N/A
Average 2.07 1.91
Source: Company data, ING
_
See back of report for important disclosures and disclaimer
40
Biotech valuation December 2004
Fig 36 Valuations of global biology services companies
Company Cur
r
Share price Mkt cap EV Sales (lc m) EV/sales (x)
(lc) (lc m) (lc m) 2004F 2005F 2004F 2005F
Curagen US$ 6.01 301.3 170.6 6.3 9.0 27.13 18.95
Cytomyx £ 16.25 6.8 13.0 6.0 7.0 2.16 1.85
DeCode Genetics US$ 6.48 353.2 436.7 48.4 52.2 9.02 8.37
Exelixis US$ 8.69 650.1 376.6 54.7 68.4 6.89 5.51
Genaissance Pharms US$ 1.65 51.4 66.6 20.6 26.4 3.22 2.52
Gene Logic US$ 3.42 108.3 50.9 78.3 87.0 0.65 0.59
Lexicon Genetics US$ 7.32 464.3 265.2 58.7 73.1 4.52x 3.63
Maxygen US$ 10.25 362.4 122.5 20.0 18.0 6.13 6.81
Pharmagene £ 40.5 21.5 6.4 5.3 9.9 1.2 0.64
Average 6.77
Source: ING
_
Tools companies
Tools companies are, like profitable companies, somewhat easier to value in view of
the fact that they have tangible earnings streams, a history of earnings and often
reasonably predictable growth profiles. There are essentially two approaches to take.
For low-growth commodity companies, we would use a combination of sales-based
(EV/sales) and earnings-based (EV/EBITDA, EV/EBIT or PE ratios) and apply
appropriate comparator multiples in each case. For high-growth companies, we would
also examine PEG ratios.
Fig 37 Peer group comparison
Bloomberg EV/EV/PER EPS
Company Ticker Share price Mkt cap sales EBITDA 2004F 2005F growth PEG
(US$) (US$m) (x) (x) (x) (x) (%) (x)
Invitrogen IVGN US 61.2 3,136.3 4.1 15.5 21.0 18.3 16.4 1.3
Techne TECH US 38.0 1,569.5 8.7 16.0 25.6 22.3 11.
5
2.2
Bio-Rad Laboratories BIO US 57.1 1,472.6 1.6 8.3 17.6 15.1 10.
0
1.8
Affymetrix AFFX US 34.4 2,097.0 6.7 40.6 49.1 31.6 41.
0
1.2
Applera Appd.Bios.ABI US 20.8 4,075.4 2.1 10.8 21.4 19.2 10.
5
2.0
Waters WAT US 47.5 5,739.9 5.9 15.9 26.5 22.7 16.3 1.6
Millipore MIL US 50.2 2,496.4 3.3 15.9 22.8 19.9 13.
7
1.7
Mean 4.0 15.2 23.1 18.5 20.
0
1.6
Qiagen QGENF US 11.2 1,639.2 4.7 16.4 31.3 25.9 19.
0
1.7
Prem/(discount) (%) 5.0
Source: Bloomberg, ING estimates
_
See back of report for important disclosures and disclaimer
41
Biotech valuation December 2004
ING’s product NPV calculator
Our NPV ready reckoner tables, or wheel, are a simple way for estimating product
NPVs based on our standardised NPV methodology (please contact us if you would
like one of our NPV wheels). Investors must be aware though that it only provides an
estimate of a product’s likely value and ultimately this cannot in any way replace the
more sophisticated approaches described earlier in this report.
To use the wheel, simply select the product’s current phase of development, together
with its likely or expected peak sales. Line up the inner wheel to the relevant sales
number for that stage of development and simply read off the NPV value assuming an
appropriate royalty rate. Although we have labelled the wheel in US dollars, it is
actually currency independent.
Investors should note that the NPV values in the tables and wheel have been included
on the basis of a number of assumptions, some of which may not be valid in specific
instances, and are intended merely as a guideline. The assumptions are as follow.
1. Time to launch is based on the basis that the product has just entered the phase
of development and the time to market is estimated as shown in Figure 38.
Fig 38 Development timeline assumptions
Clinical phase Years to market
Phase I 5 years
Phase II 4 years
Phase III 3 years
Registration 1 years
Source: ING
_
Obviously many trials run on different time scales and so an approximation to the NPV
derived from the wheel can be made by additional discounting of 13% per annum.
2. No milestones or upfront payments have been included in calculating the NPVs,
which we would normally do having analysed the type of product under review and
then utilising previous benchmark licensing deals.
3. We have used our standard probabilities for 10% success for Phase I, 30% for
Phase II and 70% for Phase III. There are obviously a number of situations where
it is appropriate to reduce the probability of success, eg, where a product is in
Phase III but has no proof-of-principle data in man – in this case, we would use
50%. All an investor need do in this case is read off the standard NPV for a Phase
III product and adjust the NPV by multiplying by 50/70 to arrive at the new NPV.
We hope you find the wheel a useful tool for an ‘at a glance’ product valuation. Happy
valuing!
See back of report for important disclosures and disclaimer
42
Biotech valuation December 2004
Appendix I
Ready reckoner NPV tables
The following ‘ready reckoner’ tables provide approximate NPVs for products at
various stages of development, using the standardised assumptions outlined earlier
(with no milestones payable). Note, these are merely intended as a guideline.
Fig 39 Standardised NPVs for candidates in early-stage R&D
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50 0.09 0.17 0.19 0.21 0.23 0.24 0.26 0.28 0.29 0.31 0.33 0.35 0.43 0.52
100 0.17 0.35 0.38 0.42 0.45 0.49 0.52 0.56 0.59 0.62 0.66 0.69 0.87 1.04
150 0.26 0.52 0.57 0.62 0.68 0.73 0.78 0.83 0.88 0.94 0.99 1.04 1.3 1.56
200 0.35 0.69 0.76 0.83 0.9 0.97 1.04 1.11 1.18 1.25 1.32 1.39 1.73 2.08
250 0.43 0.87 0.95 1.04 1.13 1.21 1.3 1.39 1.47 1.56 1.65 1.73 2.17 2.6
300 0.52 1.04 1.14 1.25 1.35 1.46 1.56 1.67 1.77 1.87 1.98 2.08 2.6 3.12
350 0.61 1.21 1.34 1.46 1.58 1.7 1.82 1.94 2.06 2.19 2.31 2.43 3.04 3.64
400 0.69 1.39 1.53 1.67 1.8 1.94 2.08 2.22 2.36 2.5 2.64 2.78 3.47 4.16
450 0.78 1.56 1.72 1.87 2.03 2.19 2.34 2.5 2.65 2.81 2.97 3.12 3.9 4.68
500 0.87 1.73 1.91 2.08 2.25 2.43 2.6 2.78 2.95 3.12 3.3 3.47 4.34 5.2
550 0.95 1.91 2.1 2.29 2.48 2.67 2.86 3.05 3.24 3.43 3.63 3.82 4.77 5.72
600 1.04 2.08 2.29 2.5 2.71 2.91 3.12 3.33 3.54 3.75 3.95 4.16 5.2 6.24
650 1.13 2.25 2.48 2.71 2.93 3.16 3.38 3.61 3.83 4.06 4.28 4.51 5.64 6.76
700 1.21 2.43 2.67 2.91 3.16 3.4 3.64 3.89 4.13 4.37 4.61 4.86 6.07 7.28
750 1.3 2.6 2.86 3.12 3.38 3.64 3.9 4.16 4.42 4.68 4.94 5.2 6.5 7.81
800 1.39 2.78 3.05 3.33 3.61 3.89 4.16 4.44 4.72 5 5.27 5.55 6.94 8.33
850 1.47 2.95 3.24 3.54 3.83 4.13 4.42 4.72 5.01 5.31 5.6 5.9 7.37 8.85
900 1.56 3.12 3.43 3.75 4.06 4.37 4.68 5 5.31 5.62 5.93 6.24 7.81 9.37
950 1.65 3.3 3.63 3.95 4.28 4.61 4.94 5.27 5.6 5.93 6.26 6.59 8.24 9.89
Peak sales (US$m)
1000 1.73 3.47 3.82 4.16 4.51 4.86 5.2 5.55 5.9 6.24 6.59 6.94 8.67 10.41
Source: ING
_
Fig 40 Standardised NPVs for products in pre-clinical development
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50 0.39 0.79 0.87 0.95 1.02 1.1 1.18 1.26 1.34 1.42 1.5 1.58 1.97 2.37
100 0.79 1.58 1.73 1.89 2.05 2.21 2.37 2.52 2.68 2.84 3 3.15 3.94 4.73
150 1.18 2.37 2.6 2.84 3.07 3.31 3.55 3.78 4.02 4.26 4.49 4.73 5.91 7.1
200 1.58 3.15 3.47 3.78 4.1 4.42 4.73 5.05 5.36 5.68 5.99 6.31 7.88 9.46
250 1.97 3.94 4.34 4.73 5.12 5.52 5.91 6.31 6.7 7.1 7.49 7.88 9.86 11.83
300 2.37 4.73 5.2 5.68 6.15 6.62 7.1 7.57 8.04 8.51 8.99 9.46 11.83 14.19
350 2.76 5.52 6.07 6.62 7.17 7.73 8.28 8.83 9.38 9.93 10.49 11.04 13.8 16.56
400 3.15 6.31 6.94 7.57 8.2 8.83 9.46 10.09 10.72 11.35 11.98 12.61 15.77 18.92
450 3.55 7.1 7.81 8.51 9.22 9.93 10.64 11.35 12.06 12.77 13.48 14.19 17.74 21.29
500 3.94 7.88 8.67 9.46 10.25 11.04 11.83 12.61 13.4 14.19 14.98 15.77 19.71 23.65
550 4.34 8.67 9.54 10.41 11.27 12.14 13.01 13.88 14.74 15.61 16.48 17.34 21.68 26.02
600 4.73 9.46 10.41 11.35 12.3 13.25 14.19 15.14 16.08 17.03 17.98 18.92 23.65 28.38
650 5.12 10.25 11.27 12.3 13.32 14.35 15.37 16.4 17.42 18.45 19.47 20.5 25.62 30.75
700 5.52 11.04 12.14 13.25 14.35 15.45 16.56 17.66 18.76 19.87 20.97 22.08 27.59 33.11
750 5.91 11.83 13.01 14.19 15.37 16.56 17.74 18.92 20.1 21.29 22.47 23.65 29.57 35.48
800 6.31 12.61 13.88 15.14 16.4 17.66 18.92 20.18 21.44 22.71 23.97 25.23 31.54 37.84
850 6.7 13.4 14.74 16.08 17.42 18.76 20.1 21.44 22.78 24.13 25.47 26.81 33.51 40.21
900 7.1 14.19 15.61 17.03 18.45 19.87 21.29 22.71 24.13 25.54 26.96 28.38 35.48 42.57
950 7.49 14.98 16.48 17.98 19.47 20.97 22.47 23.97 25.47 26.96 28.46 29.96 37.45 44.94
Peak sales (US$m)
1000 7.88 15.77 17.34 18.92 20.5 22.08 23.65 25.23 26.81 28.38 29.96 31.54 39.42 47.3
Source: ING
_
See back of report for important disclosures and disclaimer
43
Biotech valuation December 2004
Fig 41 Standardised NPVs for products in Phase I development
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50 0.79 1.58 1.73 1.89 2.05 2.21 2.37 2.52 2.68 2.84 3 3.15 3.94 4.73
100 1.58 3.15 3.47 3.78 4.1 4.42 4.73 5.05 5.36 5.68 5.99 6.31 7.88 9.46
150 2.37 4.73 5.2 5.68 6.15 6.62 7.1 7.57 8.04 8.51 8.99 9.46 11.83 14.19
200 3.15 6.31 6.94 7.57 8.2 8.83 9.46 10.09 10.72 11.35 11.98 12.61 15.77 18.92
250 3.94 7.88 8.67 9.46 10.25 11.04 11.83 12.61 13.4 14.19 14.98 15.77 19.71 23.65
300 4.73 9.46 10.41 11.35 12.3 13.25 14.19 15.14 16.08 17.03 17.98 18.92 23.65 28.38
350 5.52 11.04 12.14 13.25 14.35 15.45 16.56 17.66 18.76 19.87 20.97 22.08 27.59 33.11
400 6.31 12.61 13.88 15.14 16.4 17.66 18.92 20.18 21.44 22.71 23.97 25.23 31.54 37.84
450 7.1 14.19 15.61 17.03 18.45 19.87 21.29 22.71 24.13 25.54 26.96 28.38 35.48 42.57
500 7.88 15.77 17.34 18.92 20.5 22.08 23.65 25.23 26.81 28.38 29.96 31.54 39.42 47.3
550 8.67 17.34 19.08 20.81 22.55 24.28 26.02 27.75 29.49 31.22 32.96 34.69 43.36 52.03
600 9.46 18.92 20.81 22.71 24.6 26.49 28.38 30.27 32.17 34.06 35.95 37.84 47.3 56.76
650 10.25 20.5 22.55 24.6 26.65 28.7 30.75 32.8 34.85 36.9 38.95 41 51.25 61.5
700 11.04 22.08 24.28 26.49 28.7 30.91 33.11 35.32 37.53 39.74 41.94 44.15 55.19 66.23
750 11.83 23.65 26.02 28.38 30.75 33.11 35.48 37.84 40.21 42.57 44.94 47.3 59.13 70.96
800 12.61 25.23 27.75 30.27 32.8 35.32 37.84 40.37 42.89 45.41 47.93 50.46 63.07 75.69
850 13.4 26.81 29.49 32.17 34.85 37.53 40.21 42.89 45.57 48.25 50.93 53.61 67.01 80.42
900 14.19 28.38 31.22 34.06 36.9 39.74 42.57 45.41 48.25 51.09 53.93 56.76 70.96 85.15
950 14.98 29.96 32.96 35.95 38.95 41.94 44.94 47.93 50.93 53.93 56.92 59.92 74.9 89.88
Peak sales (US$m)
1000 15.77 31.54 34.69 37.84 41 44.15 47.3 50.46 53.61 56.76 59.92 63.07 78.84 94.61
Source: ING
_
Fig 42 Standardised NPVs for products in Phase II development
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50 2.67 5.35 5.88 6.41 6.95 7.48 8.02 8.55 9.09 9.62 10.16 10.69 13.36 16.04
100 5.35 10.69 11.76 12.83 13.9 14.97 16.04 17.11 18.17 19.24 20.31 21.38 26.73 32.07
150 8.02 16.04 17.64 19.24 20.85 22.45 24.05 25.66 27.26 28.86 30.47 32.07 40.09 48.11
200 10.69 21.38 23.52 25.66 27.8 29.93 32.07 34.21 36.35 38.49 40.62 42.76 53.45 64.14
250 13.36 26.73 29.4 32.07 34.74 37.42 40.09 42.76 45.44 48.11 50.78 53.45 66.82 80.18
300 16.04 32.07 35.28 38.49 41.69 44.9 48.11 51.32 54.52 57.73 60.94 64.14 80.18 96.22
350 18.71 37.42 41.16 44.9 48.64 52.38 56.13 59.87 63.61 67.35 71.09 74.84 93.54 112.25
400 21.38 42.76 47.04 51.32 55.59 59.87 64.14 68.42 72.7 76.97 81.25 85.53 106.91 128.29
450 24.05 48.11 52.92 57.73 62.54 67.35 72.16 76.97 81.78 86.59 91.41 96.22 120.27 144.32
500 26.73 53.45 58.8 64.14 69.49 74.84 80.18 85.53 90.87 96.22 101.56 106.91 133.63 160.36
550 29.4 58.8 64.68 70.56 76.44 82.32 88.2 94.08 99.96 105.84 111.72 117.6 147 176.4
600 32.07 64.14 70.56 76.97 83.39 89.8 96.22 102.63 109.05 115.46 121.87 128.29 160.36 192.43
650 34.74 69.49 76.44 83.39 90.34 97.29 104.23 111.18 118.13 125.08 132.03 138.98 173.72 208.47
700 37.42 74.84 82.32 89.8 97.29 104.77 112.25 119.74 127.22 134.7 142.19 149.67 187.09 224.51
750 40.09 80.18 88.2 96.22 104.23 112.25 120.27 128.29 136.31 144.32 152.34 160.36 200.45 240.54
800 42.76 85.53 94.08 102.63 111.18 119.74 128.29 136.84 145.39 153.95 162.5 171.05 213.81 256.58
850 45.44 90.87 99.96 109.05 118.13 127.22 136.31 145.39 154.48 163.57 172.66 181.74 227.18 272.61
900 48.11 96.22 105.84 115.46 125.08 134.7 144.32 153.95 163.57 173.19 182.81 192.43 240.54 288.65
950 50.78 101.56 111.72 121.87 132.03 142.19 152.34 162.5 172.66 182.81 192.97 203.12 253.9 304.69
Peak sales (US$m)
1000 53.45 106.91 117.6 128.29 138.98 149.67 160.36 171.05 181.74 192.43 203.12 213.81 267.27 320.72
Source: ING
_
See back of report for important disclosures and disclaimer
44
Biotech valuation December 2004
Fig 43 Standardised NPVs for products in Phase III development
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50 7.05 14.09 15.5 16.91 18.32 19.73 21.14 22.55 23.96 25.37 26.78 28.19 35.23 42.28
100 14.09 28.19 31.01 33.83 36.64 39.46 42.28 45.1 47.92 50.74 53.56 56.38 70.47 84.56
150 21.14 42.28 46.51 50.74 54.97 59.19 63.42 67.65 71.88 76.11 80.34 84.56 105.7 126.85
200 28.19 56.38 62.01 67.65 73.29 78.93 84.56 90.2 95.84 101.48 107.11 112.75 140.94 169.13
250 35.23 70.47 77.52 84.56 91.61 98.66 105.7 112.75 119.8 126.85 133.89 140.94 176.17 211.41
300 42.28 84.56 93.02 101.48 109.93 118.39 126.85 135.3 143.76 152.21 160.67 169.13 211.41 253.69
350 49.33 98.66 108.52 118.39 128.25 138.12 147.99 157.85 167.72 177.58 187.45 197.32 246.64 295.97
400 56.38 112.75 124.03 135.3 146.58 157.85 169.13 180.4 191.68 202.95 214.23 225.5 281.88 338.25
450 63.42 126.85 139.53 152.21 164.9 177.58 190.27 202.95 215.64 228.32 241.01 253.69 317.11 380.54
500 70.47 140.94 155.03 169.13 183.22 197.32 211.41 225.5 239.6 253.69 267.78 281.88 352.35 422.82
550 77.52 155.03 170.54 186.04 201.54 217.05 232.55 248.05 263.56 279.06 294.56 310.07 387.58 465.1
600 84.56 169.13 186.04 202.95 219.87 236.78 253.69 270.6 287.52 304.43 321.34 338.25 422.82 507.38
650 91.61 183.22 201.54 219.87 238.19 256.51 274.83 293.15 311.48 329.8 348.12 366.44 458.05 549.66
700 98.66 197.32 217.05 236.78 256.51 276.24 295.97 315.7 335.44 355.17 374.9 394.63 493.29 591.95
750 105.7 211.41 232.55 253.69 274.83 295.97 317.11 338.25 359.4 380.54 401.68 422.82 528.52 634.23
800 112.75 225.5 248.05 270.6 293.15 315.7 338.25 360.8 383.36 405.91 428.46 451.01 563.76 676.51
850 119.8 239.6 263.56 287.52 311.48 335.44 359.4 383.36 407.31 431.27 455.23 479.19 598.99 718.79
900 126.85 253.69 279.06 304.43 329.8 355.17 380.54 405.91 431.27 456.64 482.01 507.38 634.23 761.07
950 133.89 267.78 294.56 321.34 348.12 374.9 401.68 428.46 455.23 482.01 508.79 535.57 669.46 803.35
Peak sales (US$m)
1000 140.94 281.88 310.07 338.25 366.44 394.63 422.82 451.01 479.19 507.38 535.57 563.76 704.7 845.64
Source: ING
_
Fig 44 Standardised NPVs for products in registration
NPV (US$m) 5% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 25% 30%
50
11.57 23.14 25.45 27.77 30.08 32.39 34.71 37.02 39.34 41.65 43.96 46.28 57.85 69.42
100
23.14 46.28 50.90 55.53 60.16 64.79 69.42 74.04 78.67 83.30 87.93 92.55 115.69 138.83
150
34.71 69.42 76.36 83.30 90.24 97.18 104.12 111.06 118.01 124.95 131.89 138.83 173.54 208.25
200
46.28 92.55 101.81 111.06 120.32 129.58 138.83 148.09 157.34 166.60 175.85 185.11 231.38 277.66
250
57.85 115.69 127.26 138.83 150.40 161.97 173.54 185.11 196.68 208.25 219.82 231.38 289.23 347.08
300
69.42 138.83 152.71 166.60 180.48 194.36 208.25 222.13 236.01 249.89 263.78 277.66 347.08 416.49
350
80.98 161.97 178.17 194.36 210.56 226.76 242.95 259.15 275.35 291.54 307.74 323.94 404.92 485.91
400
92.55 185.11 203.62 222.13 240.64 259.15 277.66 296.17 314.68 333.19 351.70 370.21 462.77 555.32
450
104.12 208.25 229.07 249.89 270.72 291.54 312.37 333.19 354.02 374.84 395.67 416.49 520.61 624.74
500
115.69 231.38 254.52 277.66 300.80 323.94 347.08 370.21 393.35 416.49 439.63 462.77 578.46 694.15
550
127.26 254.52 279.97 305.43 330.88 356.33 381.78 407.24 432.69 458.14 483.59 509.05 636.31 763.57
600
138.83 277.66 305.43 333.19 360.96 388.73 416.49 444.26 472.02 499.79 527.56 555.32 694.15 832.98
650
150.40 300.80 330.88 360.96 391.04 421.12 451.20 481.28 511.36 541.44 571.52 601.60 752.00 902.40
700
161.97 323.94 356.33 388.73 421.12 453.51 485.91 518.30 550.69 583.09 615.48 647.88 809.84 971.81
750
173.54 347.08 381.78 416.49 451.20 485.91 520.61 555.32 590.03 624.74 659.45 694.15 867.69 1041.23
800
185.11 370.21 407.24 444.26 481.28 518.30 555.32 592.34 629.37 666.39 703.41 740.43 925.54 1110.64
850
196.68 393.35 432.69 472.02 511.36 550.69 590.03 629.37 668.70 708.04 747.37 786.71 983.38 1180.06
900
208.25 416.49 458.14 499.79 541.44 583.09 624.74 666.39 708.04 749.68 791.33 832.98 1041.23 1249.47
950
219.82 439.63 483.59 527.56 571.52 615.48 659.45 703.41 747.37 791.33 835.30 879.26 1099.08 1318.89
Peak sales (US$m)
1000
231.38 462.77 509.05 555.32 601.60 647.88 694.15 740.43 786.71 832.98 879.26 925.54 1156.92 1388.31
Source: ING
_
See back of report for important disclosures and disclaimer
45
Biotech valuation December 2004
Appendix II
Rule of thumb metrics
In practice, the pharmaceutical industry employs several rules of thumb in valuing
products. Some of these are listed below.
Figure 45 lists the average of the price paid for marketed products at different stages
of the product lifecycle as a multiple of prior-year revenues:
Fig 45 Standard M&A multiples
Product type Standard M&A multiple (x)
High-growth product 9-13
Product at or near-peak sales 3-6
Product with declining sales 2-3
Product going off-patent 1-2
Source: ING
_
Figure 46 shows the standardised NPV of a newly launched product as a multiple of
peak sakes, at a variety of given royalty rates, according to our standard methodology.
Fig 46 Standard M&A multiples
Royalty (%) Multiple of peak sales
15 0.87
16 0.93
17 0.99
18 1.05
19 1.1
20 1.16
21 1.22
22 1.28
23 1.34
24 1.39
25 1.45
26 1.51
27 1.57
28 1.63
29 1.68
30 1.74
35 2.03
40 2.32
45 2.61
50 2.91
Source: ING
_
See back of report for important disclosures and disclaimer
46
Biotech valuation December 2004
Pharmaceuticals team
Research
Max Herrmann
44 20 7767 5873
max.herrmann@uk.ing.com
London
Dr Sally Bennett
44 20 7767 5852
sally.bennett@uk.ing.com
London
Richard Parkes PhD
44 20 7767 5974
richard.parkes@uk.ing.com
London
Tim Race
44 20 7767 6541
tim.race@uk.ing.com
London
Specialist sales
Danny Allessie
31 20 563 80 85
danny.allessie@ingbank.com
Amsterdam
Dirk Marckx
32 2 557 13 72
dirk.marckx@ing.be
Brussels
Sales desks
Amsterdam
31 20 563 80 80
Brussels
32 2 557 13 70
Edinburgh
44 131 527 3000
Geneva
41 22 593 80 50
London
44 20 7767 8954
Madrid
43 91 789 00 10
Milan
39 02 89629 3660
Paris
33 1 55 68 45 00
ii
ii
i
l
EQUITY MARKETS
PharmaceuticalsWestern Europe
Dr Sally Bennett
(44 20) 7767 5852
sally.bennett@uk.ing.comRichard Parkes PhD
(44 20) 7767 5974
richard.parkes@uk.ing.com
Max Herrmann
(44 20) 7767 5873
max.herrmann@uk.ing.com
Biotech valuationAn investor’s guide December 2004
PLEASE SEE THE IMPORTANT DISCLAIMER, COMPANY DISCLOSURES AND ANALYST CERTIFICATION ON THE LAST PAGE OF THIS REPORT
ING Financial MarketsAMSTERDAMTel: 31 20 563 87 98
BRUSSELSTel: 32 2 557 10 26
LONDONTel: 44 20 7767 1000
NEW YORKTel: 1 646 424 6000
SINGAPORETel: 65 6535 3688
BratislavaTel: 421 2 5934 61 11BucharestTel: 40 21 222 1600BudapestTel: 36 1 268 0140Buenos AiresTel: 54 11 4310 4700DublinTel: 353 1 638 4000
EdinburghTel: 44 131 527 3000GenevaTel: 41 22 593 8050Hong KongTel: 852 2848 8488IstanbulTel: 90 212 258 8770KievTel: 380 44 230 3030
MadridTel:34 91 789 0030ManilaTel: 632 840 8888Mexico CityTel: 52 55 5258 2000MilanTel: 39 02 89629 3660MoscowTel: 7095 755 5400
ParisTel: 33 1 56 39 31 41PragueTel: 420 2 5747 1111SantiagoTel: 562 452 2700Sao PauloTel: 55 11 4504 6000SeoulTel: 822 317 1500
ShanghaiTel: 86 21 6841 3355SofiaTel: 359 2 917 6400TaipeiTel: 886 2 2734 7500TokyoTel: 813 5210 1500WarsawTel: 48 22 820 5018
Important Company Disclosures
The following designations [a-i] next to a company covered in this publication highlight that one or more
members of ING Group:
[a]holds 1% or more of the equity shares of the company (as at the end of the month preceding this publication).
[b]has lead managed or co-lead managed a public offering of the securities of the company in the last 12 months.
[c]has received compensation for investment banking services from the company within the last 12 months.
[
d
]
expects to receive or intends to seek compensation for investment banking services from the company in the
next 3 months.
[e]makes a market in the company’s securities in the US via ING Financial Markets LLC.
[
f
]
is a liquidity provider, or acts as designated sponsor or market maker, for the company on a German, French or
Dutch stock exchange.
[g]has lead managed or co-lead managed a public offering of the company's securities in the last 5 years.
[
h
]
has a member of its board of directors or su
p
ervisor
y
board or senior officer on the com
p
an
y
’s board of directors
or supervisory board.
[i] holds a net short position of 1% or more of the share capital, calculated in accordance with German law.
In addition, ING Group trades in the shares of the company/ies covered in this publication.Disclosures [a]-[e] are required specifically but not exclusively by US regulations.Analyst Certification and Disclaimer
The views ex
p
ressed in this re
p
ort accuratel
y
reflect the
p
ersonal views of the anal
y
st
(
s
)
about the sub
j
ect
securities or issuers and no part of the compensation of the analyst(s) was, is, or will be directly or indirectly
related to the inclusion of specific recommendations or views in this report.This publication has been prepared on behalf of ING (being for this purpose the wholesale and investment banking business of ING Bank NV and
certain of its subsidiary companies) solely for the information of its clients. ING forms part of ING Group (being for this purpose ING Groep NV and its
subsidiary and affiliated companies). It is not investment advice or an offer or solicitation for the purchase or sale of any financial instrument. While
reasonable care has been taken to ensure that the information contained herein is not untrue or misleading at the time of publication, ING makes no
representation that it is accurate or complete. The information contained herein is subject to change without notice. ING Group and any of its officers,
employees, related and discretionary accounts may, to the extent not disclosed above and to the extent permitted by law, have long or short positions
or may otherwise be interested in any transactions or investments (including derivatives) referred to in this publication. In addition, ING Group may
provide banking, insurance or asset management services for, or solicit such business from, any company referred to in this publication. Neither ING
nor any of its officers or employees accepts any liability for any direct or consequential loss arising from any use of this publication or its contents.
Copyright and database rights protection exists in this publication and it may not be reproduced, distributed or published by any person for any
purpose without the prior express consent of ING. All rights are reserved.
Any investments referred to herein may involve significant risk, are not necessarily available in all jurisdictions, may be illiquid and may not be
suitable for all investors. The value of, or income from, any investments referred to herein may fluctuate and/or be affected by changes in exchange
rates. Past performance is not indicative of future results. Investors should make their own investment decisions without relying on this publication.
Only investors with sufficient knowledge and experience in financial matters to evaluate the merits and risks should consider an investment in any
issuer or market discussed herein and other persons should not take any action on the basis of this publication. This publication is issued:1) in the
United Kingdom only to persons described in Articles 19, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order
2001 and is not intended to be distributed, directly or indirectly, to any other class of persons (including private investors); 2) in Italy only to persons
described in Article No. 31 of Consob Regulation No. 11522/98.Clients should contact analysts at, and execute transactions through, an ING entity
in their home jurisdiction unless governing law permits otherwise.
ING Bank N.V., London branch is authorised by the Dutch Central Bank and regulated by the Financial Services Authority for the conduct of UK
business. It is incorporated in the Netherlands and its London branch is registered in the UK (number BR000341) at 60 London Wall, London EC2M
5TQ. ING Financial Markets LLC, which is a member of the NYSE, NASD and SIPC and part of ING, has accepted responsibility for the distribution
of this report in the United States under applicable requirements.This document is classified as UK Inde
p
endent Research as defined in the ING Bank London Branch Research
p
olic
y
drawn up in accordance with FSA Rule COB 7.16.5R (2).
EQ_UK IND
Additional information is available on request