| |
PDF version
|
|
|
Setting
commercial terms
Introduction
Faced with the massive challenge of attempting to ensure a
steady and strong growth in sales and profit from a business
based on a small number of products which can fluctuate wildly
in sales in the space of a few years – or in the case
of products in R&D not even achieve any sales at all –
pharmaceutical industry management increasingly uses the tool
of licensing to provide itself with a more secure range of
products to sell. This requires companies to leave a proportion
of the profit from the products acquired this way with the
originator, but that is a relatively small price to pay for
gaining access to something that can rescue the manager’s
business – particularly since the margins available
on pharmaceutical products leave plenty of room for both to
benefit.
This strategic approach to the product pipeline problem has
grown in importance over the last decade or so, and now very
few managements, however proud they are of their own R&D,
would turn their nose up at a chance to gain access to an
interesting and relevant product if it came their way. It
is common to find large companies with a significant proportion
of their R&D budgets dedicated to products that came to
them in this way. In 1999 Pharmacia & Upjohn reported
that their spend on such products had gone from 4% of total
R&D expenditure in 1995 to 21% in 1999. In the same year
Eli Lilly reported a spend ratio of 20% on in-licensed products.
Over the last 11 years the most prolific deal-maker (Roche
and its affiliates) has signed 129 agreements with biotechnology
companies alone according to analysts at Recombinant Capital.
Currently about 30-50 new deals are reported monthly, each
with its own special features. Taken together these represent
a bewildering array of deal types and payments, transacted
at many different stages in the development of the business
asset involved, requiring very different activities from the
contracting partners, conferring all sorts of different rights
on the partner, with some representing total valuations of
a couple of million dollars, while others involve transactions
valued at tens or even hundreds of millions of dollars.
How are such deal terms set? How can a company which owns
what it believes to be a valuable piece of intellectual property
know where to pitch its initial proposal to a potential partner,
and decide on a deal structure which optimises the value inherent
in its IP while being acceptable to its negotiating partner?
This paper sets out to summarise some of the main factors
that must be borne in mind in preparing an out-license strategy
for a product which is still in R&D – the most common
kind of transaction, and the most problematical, since there
is no concrete sales or product approval record on which to
base negotiations.
Strategic requirement
Given the vast range of potential deal types and terms, it
is vital that a company approaching out-license should decide
in advance just what it wants to achieve through the deal,
so that negotiations can steer the debate in the most favourable
direction for the IP owner. At first sight this strategic
requirement may seem obvious – to gain revenue from
the IP. But there is always more to it than that.
When is this revenue required? Is there an urgent need for
short term cash, or will a small initial amount do, followed
by the potential for a larger amount later? It is almost always
the case that obtaining money rapidly reduces the amount that
could be earned in total. On the other hand, getting cash
immediately at least means that something will be earned,
whereas a wait for the product to move through more stages
of development brings with it the risk that the project will
fail and nothing more will be earned at all. This decision
involves an assessment of the risk profile of the IP owner
– how much risk is the company prepared to bear in order
to get the larger money that could came later? Or is the need
for cash now so great that all considerations of taking a
greater risk to gain a later greater return are irrelevant?
How confident is the company that its product will really
prove to have the wonderful characteristics that it is going
to describe to potential partners – should it take what
it can get while it is able to, because to delay might man
nothing comes in?
The next question is what form that revenue can have. Is
granting an equity position in your company to your licensing
partner an acceptable policy, or should you hold out for cash
only? This is likely to depend on internal policies within
your company, on the number of partners you ultimately might
expect to have and probably also on the type of partner you
expect to do a deal with – some companies bring more
benefit as equity partners than others. It also depends on
how much you want - if you are a smallish company with large
capital appreciation potential but want to hold out for straight
cash you are likely to get less revenue than you would if
you granted an equity stake.
What parts of the development and marketing process do you
need or want to remain involved in? This may be largely decided
for you by your capacity and skills base, but it is important
to remember that these can change over time. Astra Pharmaceuticals,
when a small Swedish company, out-licensed most of its products
for sale in the USA, including omeprazole, because it was
not at that stage equipped to market them in that country
itself. When omeprazole started to emerge as the world’s
biggest drug, it had to pay a small fortune to Merck &
Co to recover rights to that and other products initially
contracted away. Retaining any marketing rights (if your partner
will allow you to do so) will reduce the money you will receive
from a partner, both up-front and when on the market, but
may permit you access to revenues you would otherwise not
have been able to earn. Many deals include flexible terms
such as options over rights in certain countries. As well
as the marketing of your drug, there may be other areas you
want or need to retain an interest in. You may wish to keep
to yourself some of the expertise required for the development
process. You may prefer to do some further development work
because this will permit you to ask for more money from your
partner in the later stages of the product’s life because
he has been relieved of paying for early work.
Do you expect your partner to help you to achieve other strategic
objectives outside the gaining of revenue from the product?
Do you, for example, want him to set up a field force to market
the product in a particular country which you can then take
over at a pre-agreed price and point in time? Do you want
rights to any of his products in return – such “quid-pro-quo”
deals were very common several years ago and still meet strategic
needs in some cases? Do you want to be jointly active in the
development programme in order to learn from your partner?
Such options cost money, but should be considered.
These and other questions need to be answered before potential
partners are chosen and deal types are sketched out. If this
is not done, you will end up with an unsuitable deal, and
you will be much weaker in the negotiation process because
you do not know just what it is that you can yield on and
what you must insist on.
Deal types
In principle this is no limit to the kinds of deals that
may be struck with a willing partner. Indeed it is unlikely
that any two deals out of the thousands already completed
are identical.
However there is a broad pattern the underlies most of the
activity in the field, and it is important to be aware of
this, not only to help decision making but also because it
is much easier to make a deal which is recognisably comparable
with other deals already in place. Most managements tend to
be conservative, especially at senior level, where any negotiated
deal must finally be approved. It is much easier to gain approval
for a deal structure which is similar to one recently agreed
between rival companies than for one which is innovative and
adventurous. Negotiations are never conducted just with the
people sitting across the table – several layers of
management sit in the background and must equally be persuaded
of the merit of the terms.
This is not as stupid as it may appear. All those involved
in licensing of as yet unproven products are guessing at a
lot of factors. It is better to take account of accumulated
wisdom in such a vague environment than to re-invent the wheel.
Broadly the industry has settled into a pattern of terms
which involve
An up-front payment on signature
of the deal which vaguely recognises the effort that has been
put in to the product to date (but definitely is not based
on the actual amount spent – that is sunk cash and the
problem of the IP owner) and the fact of the grant of rights
in a product. As a general rule this takes the form of cash
or cash and an equity stake, and rises in value as a product
moves through the development process, as well as broadly
according to the magnitude of the product that is expected
to emerge at the end. There is also a tendency for the equity
component to be higher for products at an early stage of development.
An analysis by Recombinant Capital showed the following average
payment levels in 1997/1998: Milestone payments during the development
process to recognise that value in the product is increasing
as crucial stages of development are successfully passed.
These are usually paid in cash. The same Recombinant Capital
analysis showed total milestones payable during the period
of the average agreement to be: A contribution towards the development
costs (or complete taking over of these costs by the partner),
again generally paid in cash. Of course these only reflect
part of the story, being money paid to the other party. In
most cases the licensee will take on a lot of work in-house
which will not be evaluated in the terms of the agreement.
In 1997/1998 averages agreed were: A royalty on sales of the product
by the partner. It is rare for these to be made known publicly,
so levels can only be estimated. However there is a general
trend for them to rise from about 4-9% for an early stage
product to 10-20% or even in some cases as much as 25+% at
a late stage. Possibly an agreed supply price for the product from
the IP owner to the licensee, if the IP owner is to make the
product, which may involve a small profit element.
Occasionally a deal includes the granting
of an option to the potential licensee which gives that company
the exclusive right to by into the product at a later stage,
perhaps when more data is available.
Valuation
You will need to prepare a financial model which shows what
you and your licensing partner are likely to earn out of any
deal. Since you also need to know what your product is in
total worth in order to position it in negotiations it is
best to construct a model which does both jobs. It is worth
investing some time and effort into this process, partly because
the results are important and should be as robust as you can
make them and partly because the process of creating such
a valuation model will teach you a great deal about your product
and its market.
The preferred way of doing this is to create a simulation
model of the product in its market. Simulation is a method
of building a financial picture which takes into account the
fact that most of the assumptions you make for an early stage
product will be guesses. It permits you to use a range of
numbers to estimate any factor that is important in the model
rather than a single figure, and to associate a probability
with numbers within the range. For example you can estimate
that product launch has a 30% chance of occurring in 2005,
a 50% chance of occurring in 2006 and a 20% chance of occurring
in 2007, and that its price could be anything between $20
and $32 per month, with the most likely figure being, say,
$25. You can even make the model recognise several completely
different outcomes – for example that the product has,
say, a 40% chance of failing ever to go beyond Phase III,
in which case it will earn nothing in the market at all. From
all of this, key results such as product Net Present Values
(NPVs) can be calculated much more meaningfully than with
normal single digit, single scenario, assumptions.
To do this the model creates several thousand scenarios based
on all of these many inputs, which it will analyse for you
in a variety of ways, for example to provide an average result,
or a probability that the NPV will exceed, say, $20 million,
or fall below $3 million.
At this level, the model has just calculated the total value
of the project. It then needs to have elements added which
will reflect potential deal terms, such as milestone payments
and royalties, and to calculate both aspects of this –
what you will earn and what your partner will earn. These
figures will again be output in the form of ranges which you
can analyse in a variety of ways.
Option pricing
Recently a good deal of attention has focused on an alternative
method of evaluating products in R&D – option pricing.
In principle this is a valuable approach, which recognises
that those elements of the valuation equation which tend to
reduce a product’s value if a very traditional approach
to valuation is used (not the simulation method described
above) – uncertainty over how it will perform in the
market, and the length of time before results will be achieved
– are actually positives. However the methodology is
as yet unproven, involving assumptions which are hard to verify.
Decision tree
Once a value for the product in the market has been established,
the data generated can be used to create a simple but effective
tool to enable management to make final decisions on the nature
and timing of the optimal deal. This involves creating a decision
tree which reflects the main stages the product has to go
through, the deal terms that might be used and the value of
the product once in the market. To illustrate how this can
be done an example of a set of potential deals has been generated
based on the following assumptions:
The product in question is about to enter Phase I The costs of Phase I and II together are estimated
at $10 million, the probability that these phases will be
successfully completed is 27.5% The cost of Phase III is $30 million. The probability
of successful completion is 60%. Based on a simulation model of the product’s
potential performance in the market (as described above) it
is assumed that the NPV of the product’s performance
might be:
With a good profile (probability 25%) $500m With a neutral profile (probability 40%) $250m With a bad profile (probability 35%) $50m The owner of the IP has decided that three kinds of deal are possible Co-development and co-marketing world wide. Consequences for the company: Immediate out-licensing, in which case the company receives: Out-licensing at the end of Phase II. Consequences:
A decision tree is then constructed to
reflect these deal possibilities. The basic structure is as
follows:
The picture is completed by adding the rest of the branches
for the license now and license end Phase II options, and
including the probabilities of success at each stage, the
cost of each stage and the NPV of revenues earned.
This permits the software to calculate what each option is
worth, and what each option exposes the company to in terms
of risk, as follows:
This highly informative image provides a great deal of information
about the strategic choice that the company faces. The figures
in the boxes are the values that each part of the tree to
the right of the node by the box represents. So for example
in the co-development case the product is worth (in formal
terms its “expected value” is) $77 million after
it has succeeded in Phase III. However at present it is worth
$6.9 million, because there are several negative events that
could occur before it has a chance to reach the end of Phase
III. Overall the preferred option, based just on the value
that each represents at the present time, is the co-development
route, because it is worth $6.9 million, whereas licensing
now is worth $6.6 million and licensing at the end of Phase
II (intuitively perhaps the most beneficial option) is actually
worth a negative figure, –$1.7 million.
However before this route is finally selected it is important
to consider the strategic imperatives of the company, and
see whether these might change the decision. This is particularly
so since the current values for co-development and out-license
now are so close to each other. Examination of each sub-tree
shows that while their overall value is very similar, their
risk profiles are definitely not. In the license now case,
there is no risk of the company losing money at any stage
in the process, however even if the product is very successful
(which only has a 4% chance of occurring) it will only earn
$42 million in revenue. Its most likely outcome (probability
73%) is an income of $3 million. In the co-development case,
by contrast, there is a small possibility (again 4%) of earning
$180 million, but the most likely outcome (at 73%) is to lose
$5 million.
This presents a valuable insight into the dilemma as first
conceived. The next step would be to re-visit the deal assumptions
and test out some potentially acceptable alternative deal
structures in the decision tree. It is important to do this,
both to find the optimal deal structure and to discover the
dynamics of the situation when even small changes could change
the picture considerably. This is best done by performing
sensitivity analyses on various assumptions.
Valuing the product through simulation and then presenting
deal options in this way is clear and understandable even
to those not involved in the details of the decision-making
process. It also provides a vital back-up for the actual negotiation
sessions.
Conclusion
Arriving at acceptable licence terms is an art rather than
a science, and by far the most important influence on what
can be achieved is the history of comparable deals. However
careful analysis of what the owner of the IP wants to achieve
(as well as close consideration of what the potential partner
will be seeking), allied to a full appraisal of the scale
and dynamics of the product’s potential using simulation
modeling and decision trees, will put the owner of the product
in a strong position to obtain optimal terms from a negotiating
partner.
Top
|
