The cost of R

1
The cost of RD How much money is needed to
address the current need?

• Andrew Farlow
• Department of Economics, and Oriel College
• University of Oxford
• Médecins Sans Frontières (MSF) Campaign for
  Access to Essential Medicines 
• International Conference on Ensuring Innovation
  for Neglected Diseases
• London, United Kingdom
• 8 June 2005
• www.economics.ox.ac.uk/members/andrew.farlow

2
Its not just RD

• 42 of Africas population 300 million people
  have no access to safe water.
• Without clean water, anti-retroviral treatment
  for AIDS sufferers is not as effective, and
  formula milk cannot safely be used to prevent
  transmission of HIV from mother to child.
• Better water management can greatly reduce
  malaria mosquito breeding sites.

3
Its not just RD

• Two-thirds of all the African children who die
  under the age of five could be saved by low-cost
  treatments such as vitamin A supplements, oral
  rehydration salts and insecticide-treated
  bed-nets to combat malaria.
• A tenth of all the diseases suffered by African
  children are caused by intestinal worms
• These can be treated for 25 US cents per child
• Research on virus resistant maize for Africa
• LOADS OF EXISTING UNDERUSED TECHNOLOGIES

4
But RD is vital

• Huge range of issues
• Better drugs, e.g. artemisinin combination
  therapies (ACTs) for malaria
• Better diagnostics
• Better prevention
• Better delivery systems
• More heat-stable products
• Better
• Gates Grand Challenges
• Why should the poor not benefit from
  technological advances too?
• A matter of equity.

5
A mix of RD cost studies

• Tufts (DiMasi, Hansen, and Grabowski)
• The Price of Innovation New Estimates of Drug
  Development Costs 2003
• Assessing Claims about the cost of new drug
  development A Critique of the Public Citizen and
  TB Alliance Reports 2004
• Public Citizen
• Rx RD Myths The Case Against the Drug
  Industrys RD Scare Card 2001
• The Global Alliance for TB Drug Development
• The Economics of TB Drug Development 2001

6
Views of one side

• (Referring to Public Citizen Report) fundamental
  economic principles were ignoredtheir estimates
  using published data are deeply flawed and
  substantially understate RD cost.grossly
  technically inappropriate fallacious economic
  reasoning about the nature of investment in RD
  and an erroneous view of the corporate income
  tax Tufts 2004
• For numerous reasons, the projections in the TB
  Alliance Report cannot be appropriately compared
  to our 802 million resultOur RD estimates are
  costs per approved drug, not costs per approved
  indication Tufts 2004

7
And the other side

• But this RD scare card or canard is built
  on myths, falsehoods and misunderstandingsan
  unfounded or false, deliberately misleading
  story.
• Public Citizen referring to TUFTS

8
Average cost of a drug?

• Tufts 802m now 1.2bn in some versions.
• TB Alliance 115m-240m.
• Another angle if divide total global
  pharmaceutical spending by output of NCEs 50bn
  per NCE

9
Role of CIPIH and others

• Most of the above are studies of drug development
  costs.
• We need more data on non-drug RD too.
• Vaccines
• Innovative interventions of all sorts too
• CIPIH
• Flip side to CIPIH work is costs of RD. Better
  ways (including IP) to do RD gt lower RD costs.
• We need to gather information on RD costs and
  other factors needed for economic analysis.
  (CIPIH Open Forum 1 June 2005, Overview slide 40)

10
All studies have issues TUFTS

• Just a few issues
• 1) Sampling issues
• Those approached to join sample were not
  themselves a random sample.
• Failed to check that the voluntary nature of
  response did not create further sample bias
  toward more costly firms (Tufts found anecdotal
  evidence of this).
• Mergers may also have helped bias the sample, by
  self-selecting firms to take part in the
  direction of high cost firms.
• Trial size issuesespecially that they are
  measured as high size and high cost in Tufts
  compared to other approaches to working out costs
  of development.
• Secrecy of the underlying data.. Cant
  independently replicate the RD cost
  experiment a standard scientific requirement.

11
All studies have issues TUFTS

• 2) Skewed cost distributions, with mean greater
  than median, but the mean used in calculations.
• Since successful trials were used to price the
  unsuccesful trials, and since we know that firms
  strategically spend more on trials that are
  likely to succeed, we record a higher cost for
  failed trials than those trials did on average
  actually cost.
• Ideally we would like the size-distribution of
  surviving and non-surviving trials at each phase
  and calculate out of pocket cost on that basis.
• Another problem of the secrecy.

12
All studies have issues TUFTS

• 3) Covers self-originated new molecular entities
  (NMEs). Adequate for the described task.
• But the self-originated makes them a more
  expensive sample than average.
• May be inadequate for many of the issues raised
  by current neglected drug projects.
• Hard to read into Tufts cost data what it might
  mean for, say, current neglected diseases work.
• Date a poor proxy for e.g. AIDS drugs (c.f. South
  African Competition case)
• Inadequate for debating in an informed fashion
  the relative roles of private/PPP v public
  sectors.

13
All studies have issues Public Citizen

• Just a few issues
• 1) If the point is to work out what the real
  resource costs are of developing a drug/vaccine,
  then should not adjust down to remove tax
  breaks. Should leave pre-tax.
• If want to have a debate about profitability of
  the industry, gov. subsidies, excessive pricing,
  etc. then we want to adjust for some of these
  things.
• Meanwhile real resource cost is a pre-tax
  issue.
• 2) Wrong to remove capital costs.
• Who pays capital costs, and where risk ends
  up is part of the problem and helps choose
  between RD systems. Capital costs need to be
  factored in.
• 3) Not enough time spent on worrying about TUFTS
  methodology.

14
General points

• Claim of an average cost to develop a
  drug/vaccine is nice but misleading
• Its not an exact science.
• It depends what you are looking for (cryptic
  comment sorry!).
• A wide range is to be expected.
• Mechanisms highly dependent on knowing this
  supposed average cost are to be avoided.

15
Some general themes instead

• 1) Opportunity Cost
• 2) Rent seeking behaviour
• 3) Risk and the cost of risk
• 4) Emerging economy costs of RD

16
1) There is always a budget constraint

• Massive level and range of needs there is
  always a budget constraint.
• Opportunity cost is not a dirty phrase
• The alternative you were prevented from doing
  because you spent on this project
• For any given budget for developing
  drugs/vaccines/diagnostic devices/delivery
  devices you could have achieved MORE
  drugs/vaccines/ diagnostic devices/delivery
  devices and access by avoiding high cost
  approaches.
• It should guide intertemporal policy decisions
• The overall RD cost matters (i.e. efficiency of
  underlying RD approach taken) and not the timing
  of the flow of funding per se.

17
Commission for Africa costings There is always a
budget constraint
Opportunity cost the alternative you were
prevented from doing because you spent on this
project
18
2) Minimize rent seeking to keep RD costs down

• A CASE-STUDY COST OF DRUG RD IN GENERAL
• Say, in a perfectly functioning RD scheme,
  it would cost an expected 1bn to discover the
  drug/vaccine for a disease in a way that
  maximizes the social surplus at 5bn.
• You wish to get as close to the optimal scheme
  as possible (you never make it the world is
  imperfect)
• You want to leave consumers themselves with as
  much of the social surplus as possible and not
  have it taxed or priced away from them
  especially if they are poor and the opportunity
  cost of resources is high.

19
2) Minimize rent seeking

• What happens if instead you offer all the 5bn of
  surplus to cover the RD? 

20
2) Minimize rent seeking

• Rent seeking Even very inefficient projects
  that would not survive under the perfect scheme
  will survive.
• Although their probabilities of success are poor,
  they will still attract financial backers, since
  on average they generate positive expected
  profit.
• In markets with a lot of economic rent
  possibilities, a large part of the social surplus
  is dissipated
• Excessive marketing (though we need to allow
  for useful marketing).
• Excessive me toos (again allow for useful
  me toos).
• High cost forms of RD, including large sample
  sizes to prove small differences in therapeutic
  value, trials that include marketing reasons (big
  industry literature on this), etc.
• Inefficient use of IP.
• Inefficiency is compounded by the deadweight loss
  of having to raise the 5bn if it is raised
  through, say, taxation.

21
2) Minimize rent seeking

• Fair return on industry capital entirely
  consistent with inefficiencies in end product
  and/or intermediate product markets.
• (Scherer example) If for example, patents turn
  out to be inefficiently long, rather than
  perennial excess returns being earned, most of
  the potential private benefit is dissipated in
  rent seeking behaviour transforming these rents
  into costs even if there is little or no social
  benefit.
• This drives profits to the point where price
  equals average costs, where fair return is
  achieved.
• High measured RD cost figures are endogenous to
  market power and industry structure as much as
  being dependent on any fundamental notion of
  RD costs.

22
2) Minimize rent seekinglessons

• Other thoughts on this We want reward to be
  linked as much as possible to the true costs and
  difficulty of developing products for each
  disease as well as therapeutic value,
  epidemiology, etc.
• So the optimal RD cost of any project should
  depend on a range of variables (all expected
  too!!!)
• Complexity of underlying science
• Costs of doing RD (also depending on types of
  firms encouraged)
• Epidemiology
• Production costs of the eventual product, etc.
• NOTE Not just information on the medical
  condition itself.
• Note not necessarily linear tradeoffs.

23
2) Rent seeking and neglected disease?

• The tight budget constraints of neglected disease
  research suggest that rent-seeking and forms of
  wasteful activity are likely much lower than for
  developed economy disease RD
• RD for neglected diseases especially cost
  effective.
• At the margin a spent on neglected disease RD
  is likely to be especially productive.
• RD cost studies based on developed economy
  markets are likely to ill-fit RD costs of
  neglected disease markets.
• Need for more separate neglected disease specific
  RD cost studies.

24
2) Rent seeking and neglected disease?

• The distortion away from neglected diseases is
  linked to rent seeking activity in developed
  economy markets
• Developed economy health RD problems and
  neglected disease RD problems are linked.
• Change incentives/reforms in one what happens in
  the other?

25
2) A lesson avoid rent seeking in any new RD
mechanisms

• OBSERVE When considering new mechanisms to
  support neglected disease RD, RD costs are
  lower by avoiding mechanisms that create rent
  seeking (or by thinking of ways to modify them to
  avoid rent seeking)
• E.g. An APC two-stage game.
• If lots of discretion and need to ex post adjust,
  then APC has firms competing twice - at the RD
  stage, and again, at the committee stage.. The
  second stage competition is rent seeking
• Or rent seeking happens at start of mechanism to
  reduce the number of potential players at the
  second stage.
• Any mechanism with committees (including
  Treaties) and discretion, and large sunk costs
  still to be paid at the end will encourage rent
  seeking and higher per unit RD costs.

26
2) A lesson

• OBSERVE not just emphasis on therapeutic value
• Eg. Prize-based models could not just reward by
  therapeutic outcome.
• It would ignore the differential underlying cost
  space
• Choice of prize-based scheme over alternatives
  (front loaded, open source, directed research,
  etc.) would boil down to relative efficiencies at
  achieving close to the optimal 1bn scheme.
• Similarly models that just target costs, miss out
  on the therapeutic profile.

27
3) RISK and cost of RD finance

• Tufts half the RD cost of developing a new drug
  is cost of capital i.e. reward to risk taking.
• Wrong to exclude capital/risk costs from
  commercial RD cost figures.
• Wrong to exclude capital/risk costs of
  non-commercial RD models too someone bears the
  risk
• But it is also wrong to exclude the value of
  risk-saving of such RD models!
• If PPPs/NIH activity takes away risk from later
  players, should that not be properly valued and
  ascribed too (currently it is not in RD cost
  data)?
• Debate about relevant level of capital costs and
  who bears the risk, should be part of the choice
  about how and where to do RD.

28
3) RISK and cost of RD finance

• Is RD high risk? Well, yes and no.
• Degree of individual variance of an RD project
  or of a pharmaceutical firm is not what matters
• What matters is the degree of covariance of that
  firm with the overall market.
• If financial markets are efficient, risk can be
  spread, with investors holding well-diversified
  portfolios and bearing little or no idiosyncratic
  risk.
• This is in the capital (CAPM) cost methodology
  used But this is lost from the rhetoric and PR
• The talk is always about individual financial
  gambles on technology, and never the covariance
  of these financial gambles with other financial
  gambles.

29
3) RISK and cost of RD finance

• Most of the TUFTS capital cost is the market
  rate.
• Big increase in capital costs component in Tufts
  study was not because RD got riskier per se
  (that is a separate issue) but because the stock
  market bubbled in the late 1990s
• Pulled up average cost of capital rate up
• This got applied on earlier sunk RD costs too

30
3) Equity/non-equity based RD finance

• Traditional model of drug development has been
  equity-based RD.
• PPP and others have changed this.
• What are the implications of this for the RD
  cost question?
• This is an under-considered point.
• But all mechanisms of RD, including all proposed
  new mechanisms, have different presumed financial
  mechanisms underlying themand this has
  implications for RD costs.

31
3) Equity/non-equity based RD finance

• Some thoughts on the role of equity finance in
  drug/vaccine RD.
• Fundamental financial problem separation of
  ownership and control of firms engaged in
  research. Two affects
• A managers/scientists have a preference to
  invest in things that benefit them (a larger firm
  size, nicer offices, more staff under their
  control, higher pay, prestige projects, etc.)
• B being risk averse they wish to avoid risky
  RD.
• Normally, one would think to use more debt than
  equity finance (leveraging) to mitigate problem
  A.

32
3) But leverage is of limited use in the case
of RD-intensive firms

• 1) The knowledge asset created by RD
  investments is intangible, often contains a lot
  of know-how, is partly, if not largely,
  embedded in human capital, and is often very
  specific to the firms . for debt-holders there
  is no physical asset to secure loans (i.e..
  debt).
• 2) Servicing debt requires a stable cash flow.
  Often RD must be sustained at a certain stable
  level to be productive and it would make RD even
  more expensive if it had to compete with this
  cash flow requirement.
• 3) If bankruptcy is a possibility, managers may
  avoid variance-increasing RD projects that have
  value and that shareholders would want, leading
  to fewer long term projects.

33
3) Role of equity finance in RD

• So, the apparent solution to the first problem
  doesnt work.
• So, most pharmaceutical RD instead takes place
  in
• firms based on equity-based, external, finance
• older firms with already established cash flow
  records
• or newer firms with access to venture
  capital...(but venture capital is also
  expensive).

34
3) Role of equity finance in RD

• But this leads to a new set of problems
• One reason firms do not do certain kinds of
  research is because it is hard to communicate to
  equity-based markets the value of research (even
  if they want to be truthful since they will not
  be believed) and hence to raise the finance for
  it.
• asymmetric information and moral hazard gt extra
  gap between the private rate of return and the
  cost of capital when the innovator-investor and
  financier are different.
• Firms therefore do not invest in innovations that
  would pass the private returns hurdle.
• Short termism too Jon Horton, GSK, says firms
  like to see a return on investment by the end of
  year 3.

35
3) Role of equity finance in RD

• Also by its very nature drug RD is very long
  term. The lemons premium is higher for RD than
  for ordinary investment because the difficulty of
  separating good from bad projects when projects
  are long-term RD investments is much greater
  than with short-term low-risk projects.
• The problem is made worse by the fact that many
  firms are also reluctant to release information
  to financial markets, afraid of revealing
  information to competitors.
• Paradox Need secrecy to secure the end
  payment, but this enforces lack of sharing and
  higher cost forms of finance.

36
3) Where is this argument going?

• Are PPPs, breaking some of these information
  problems? Or not?
• Are PPPs reducing risks and enabling cheaper
  forms of finance or not?
• Are pay-as-you go ways to finance RD lower (or
  higher) cost approaches to neglected drug RD
  compared to end-to-end (i.e. all reward at the
  end)?
• How do PPPs diversify their portfolio if they are
  less equity based?
• This should instruct our choice of RD mechanism?
• E.g. 10-15 malaria vaccines entering clinical
  trials in Africa in next few years ten of them
  European. Who to direct funding to these
  developers or big industry players?

37
3) Costs of finance matter case of HIV

• HIV vaccines likely to take a minimum of 15 years
  to develop.
• If we take (at face value) 1.2bn per year of
  out-of-pocket research and trial costs needed
  (IAVA 2004).
• Replacing this flow for 15 years with a payment
  at the end, and relying heavily on biotechs and
  venture capital at the start and large
  pharmaceutical firms later. How much would it
  cost?

38
3) Costs of finance matter case of HIV

• Required rate of return to investment needs to
  capture many risks above and beyond required
  market return
• Uncertainty about ever getting a vaccine
• Uncertainty about internalizing value of research
  in a highly open iterative discovery process such
  as that required for HIV
• Risk that the mechanism collapses at any point
  between now and end
• High level of time-inconsistency risk
  (all-at-the-end APCs still contain a great deal
  of time inconsistency risk)
• High cost of venture capital for early
  discoveries (30-40 is typical VC rate we need
  to average that into the 15 year rate), with
  later rates lower
• High reputational risk for the last big firm in
  the chain.

39
3) Costs of finance matter case of HIV

• How much above normal required market rate of
  return?
• And how much to pay at the end?
• 65bn, if required nominal rate of return 15,
  real 12)
• 105bn, if required nominal rate of return 20,
  real 17)
• 165bn, if required nominal rate of return 25,
  real 22)
• This excludes the (still likely very high) costs
  after 15 years since APC would not pay out all in
  first year so as to allow follow-on vaccines.
• These are low rates of return compared to
  speculative investments that venture capital
  normally makes.
• But are they too high for this case? Or too low?
• What are the exact sizes of the risks in the case
  of HIV and how high are capital costs going to
  be?

40
HIVlooking from the other side

• Average expected horizon until repayment
• 10 year horizon each 1billion pays for about
  100m-150m of early out-of-pocket HIV research
  costs
• 15 year horizon each 1billion pays for about
  35-66m of early out-of-pocket research costs.
• All in expected terms and all very, very rough

41
HIVthe other way..

• Add in crowding out
• say, half (maybe push payments prove hard to
  remove from winners, and Russia, India and
  China cannot be barred from spoiling markets
  for products later)
• 1billion of promised HIV payment paying for
  about
• 50-80m of genuinely additional new early
  out-of-pocket private RD at 10 year horizon
• 15m-30m at 15 year horizon

42
HIVthe other way..

• Maybe this is why current levels of private
  funding are so low?
• It is claimed there is no market.
• Maybe it is the very high risk, high capital
  costs and high risks of crowding out?
• Does a 3bn fund have any impact at all?
• If it, at most, creates a few months worth of
  what IAVI says is needed, will firms bother? It
  would be too risky?

43
HIV. The dangers.

• Unfortunately, as public budget deficits prevail
  across OECD countries, there seems little
  prospect of major new public initiatives on a
  scale to make a significant difference. Harvey
  Bale, CIPIH posting CIPIH Forum, 7 Mar 2005.
• Strong pressures to trim current levels of
  funding for HIV vaccine research due to the size
  of budget deficits.
• Proposed U.S. budget, includes only 0.5 percent
  increase in overall funding for the NIH
• Substantially less than the rate of inflation
  during the past few years
• Way below the rates of funding increase of the
  past decade.

44
HIV The dangers

• Our belt is being tightened for us...the
  previous largess that was associated with all
  research, particularly HIV, is now not going to
  be a reality for the future." Dr. Anthony Fauci,
  head of the National Institute of Allergy and
  Infectious Diseases, NIAID,
• Fauci says that this tightening may well hit HIV
  vaccine research especially hard
• So, lots of hype about power of HIV APCs for
  HIVcombined with pressures to cut back funding
  for vaccine research
• But we saw just how weak an HIV APC would be..
• What are the consequences for HIV vaccine RD and
  that chances of ever getting a HIV vaccine?

45
4) The role of emerging countries
The Global Alliance for TB Drug Development The
Economics of TB Drug Development, 2001, p58.
46
4) The role of emerging countries
The Global Alliance for TB Drug Development The
Economics of TB Drug Development, 2001, p59
47
Concluding Thoughts

• Not driven by what is the cost of RD... More
  important is to work out the most efficient way
  to do RDincluding nature of financial
  instruments used to pay for research and then
  see what that costs
• Then use that to work out how much is needed