Pioneers, Imitators, and Generics

1
Pioneers, Imitators, and Generics A Simulation
Model of Schumpeterian Competition

• The authors develop a computer simulation model
  of RD competition in new product introductions
  and apply it to the pharmaceutical industry
• There are two qualitatively different types of
  RD strategies in the model pioneering and
  imitative
• Pioneering RD involves the development of a new
  family of products with the potential for very
  significant therapeutic advances and commercial
  success risky but potentially very rewarding
• Imitative RD involves investigating a known
  family of products to produce incremental
  improvements less risky but less rewarding
• A third type of strategy that involves no
  innovation is generic competition, which the
  authors consider but do not endogenize

2
The Goal Policy Impacts

• The parameters are based on data for the U.S.
  pharmaceutical industry during the 1970s
• The authors examine the sensitivity of innovation
  levels to the rate of generic competition,
  regulatory review times, and patent lives
• They are particularly interested in how changes
  in government policy influence the rate of
  innovation and the relative positions of the
  different types of firms

3
General Description of the Model

• Competition centers around investments in RD
  projects
• The firms in the model can be characterized as
  collections of ongoing projects in various stages
  of the life cycle discovery, development, market
  life, removal from the market
• Each year, the firm must make various decisions
  about its portfolio of RD projects whether to
  continue funding each one, start new ones, exit
  the industry, etc.
• Completed RD projects receive random qualities
  that determine commercial potential, and the path
  of sales follows a life cycle pattern (rising
  initially, then stable, then declining)
• The pattern of sales if subject to change as a
  result of imitative entry

4
General Description, cont.

• The earnings from successful RD projects are
  used to fund future RD
• Firms engaged in pioneering RD plow back a
  target percentage of their total earnings into
  RD, subject to several constraints
• Firms engaged in imitative RD base their RD
  expenditures on expected and realized earnings
  from RD
• They assume firms specialize and either pursue
  pioneering RD or imitative RD

5
RD and Profits

• Pioneering and imitative RD have similar
  features, but the expected value of RD costs and
  market sales are greater for pioneers
• A pioneering strategy requires a discovery phase
  that imitators do not have to perform
• Consequently, the RD process is longer and more
  costly for pioneering firms
• Pioneering RD has a lower probability of success
• Pioneering firms face a more highly skewed
  distribution of market sales

6
RD Competition Across Market Classes

• All competition takes place in the context of
  particular industry submarkets
• The allocation of RD budgets across submarkets
  is important
• They assume the probability that a pioneering
  firm launches a product in any particular
  submarket is equal across submarkets
• However, any submarket that has a major product
  breakthrough (a blockbuster) is assumed to have a
  temporary depletion of the stock of opportunities
  which prevents that class from having another
  such winner in the immediately subsequent years
• Imitators target their RD allocations to
  submarkets with recent market successes their
  new product launches are a lagged stochastic
  function of the market introduction pattern of
  the pioneers
• An imitators payoff is a function of the sales
  of the pioneer it is imitating and the number of
  competing imitative products entering the
  submarket
• Pioneers new product introductions are more
  market expanding and less redistributive than
  imitators RD

7
RD Expenditure Functions

• Pioneering firms plow back a percentage of their
  net revenues into RD, subject to constraints on
  the growth of RD and the funding of ongoing
  projects
• The pioneers operate with a target RD
  expenditure to revenue ratio
• They use the earnings targeted for RD first to
  fund ongoing projects and then to initiate new
  projects
• A constraint on RD growth is incorporated to
  avoid unrealistically sharp increases in RD from
  large increases in net revenues
• If funds allocated for RD are less than the
  amount needed for ongoing projects, no new
  projects are begun. Up to a point, existing
  projects continue to receive funding
• When the number of new blockbusters increases,
  imitators increase their targeted RD to earnings
  ratio subject to constraints on RD growth
• If the retained earnings targeted for RD drops
  below 50 of what is necessary to fund ongoing
  projects, the firm stops all RD activity

8
Policy Experiments

• They perform several policy experiments change
  patent life, for example, and observe the effects
  on pioneering RD, imitation, etc.
• In these experiments, they assume the target RD
  to sales ratios do not change when the policy
  changes
• The experiments incorporate parameter estimates
  for RD costs, technical and commercial success
  probabilities, sales revenue distributions, and
  other parameters that match the pharmaceutical
  industry during the 1970s
• The main policy variables analyzed are regulatory
  review times, patent life, and the degree of
  sales loss to generics after patent expiration

9
Base Case Parameters and Modifications

• Based on the 1970s
• 1. The average discovery phase is three years,
  development takes eight years
• 2. Only pioneering firms need discovery, so
  imitators have a longer effective patent life (11
  years compared to 8)
• 3. After patent expiration, 5 of the sales are
  lost to generics in the first year and 10 in all
  subsequent years of market life
• They compare this to more modern times, where the
  average expected loss to generics would be 50
  after patent expiration
• The impact depends critically on the patent life
  with a short patent life, rapid generic entry is
  particularly damaging to pioneering firms
• The 1984 Act extended the patent life for
  pioneers they also consider this effect, looking
  at increases of three and six years
• They also consider reductions in regulatory
  review times, which increase positive cash flow
  earlier in the life cycle than patent extensions

10
Simulation Experiments

• They assume the industry initially consists of
  twenty equal sized firms
• Ten are pioneers ten are imitators
• In year 1, each firm is already a going concern
  and owns a portfolio of drugs developed and
  marketed over the preceding 20 to 30 year period
• It is also performing RD for projects in various
  stages of the development process
• Each firm earns 57 million in the first period
  and spends 28 of this on RD. Given the projects
  in progress, there is an expected value of 6
  pioneering products and 15 imitative ones each
  year

11
After 50 Years

• After 50 years have elapsed (average results from
  10 simulations), industry net revenue has
  approximately doubled
• The average market share of the pioneers is
  significantly larger than that of the imitators
  (56 compared to 44)
• Imitators have more product introductions, but
  they have smaller average sales than pioneers
• Pioneers have higher RD
• The market share of the top four firms rises from
  20 to 42

12
Changes in Generic Competition

• Consider moving from the base case of 10 sales
  loss after patent expiration to 50
• The assumptions on effective patent life for
  pioneers and imitators, 8 and 11 years, are
  retained
• The rate of innovation and other performance
  variables are highly sensitive to this assumed
  increase in the degree of generic competition
• Industry revenues and RD are reduced by
  approximately 30 and new product introductions
  by 20
• If generics hurt pioneers more than imitators
  (60 loss for pioneers, 40 for imitators), then
  the revenues and RD of pioneers fall more
  relative to that of imitators

13
Changes in Effective Patent Life

• The 1984 Act encouraged generic entry but also
  extended patent lives for innovators
• To consider this effect, they maintain the
  assumption of a 50 post-expiration sales loss to
  generics
• They assume that patent lives increase by three
  years for pioneers and imitators
• As a result, industry RD and net revenues
  increase by over 20 relative to the case without
  patent extensions
• Approximately 75 of the loss from enhanced
  generic competition is restored by the three-year
  increase in patent life.
• These benefits are subject to diminishing
  returns a six year increase in patent life has a
  very small marginal effect relative to the three
  year increase. This is because the gains come
  late in the products life cycle when revenues
  are already low

14
Changes in Regulatory Clearance Times

• They consider reducing regulatory clearance times
  by one and two years
• They find that innovation levels are very
  sensitive to this change
• A one-year reduction increases revenues and RD
  by approximately 30
• There is a slight tendency for pioneers to
  benefit more than imitators
• There are diminishing returns in moving from one
  to two years, but these are not as severe as in
  the case of patent extensions
• The stimulative effects on innovation of a
  one-year reduction in regulatory approval times
  is equivalent to a five or six year patent
  extension
• If a firm is able to increase its cash flow early
  in the life cycle, it is able to initiate new
  projects sooner, realize returns on them sooner,
  start more new projects, etc. The cumulative
  effect is substantial
• Even if reductions in regulatory times are
  accompanied by equal reductions in patent life,
  there is still a substantial positive effect
  (regulation has no benefit in this model it is a
  pure cost)

15
Robustness

• They consider the sensitivity of their results to
  changes in the number of firms participating in
  the industry, the extent to which new product
  sales are market expanding vs. redistributive,
  and the richness of technological opportunities
• Reducing the number of firms had no effect
• If business stealing rises then revenues and
  product introductions fall (this could be
  endogenized. If more products are introduced then
  business stealing would rise they are competing
  in similar markets)
• Improving technological opportunities raises
  revenues and innovation
• They do not allow for entry into the industry