You are on page 1of 29

RESEARCH &

DEVELOPMENT
Topic 7
CH 18 C&W
Topic outline

■ 18.1. A positive analysis: Strategic


R&D
■ 18.2. market structure & incentives
for R&D
INTRODUCTION
• Innovation – the development of new
products and processes – is an important area
of industrial organisation.
• It is equally an important part of growth
theory and macroeconomics more generally.
• Joseph Schumpeter in early C20 wrote about
“creative destruction”, the process of creating
new products and processes that displace
existing ones
• Schumpeter argued that this kind of
competition is more important than price
competition at a given point in time
• CW’s chapter 18 covers a sequence of
increasingly sophisticated models of
innovation – we follow this lead
• Starting point is familiar from topic 6 –
incremental cost reduction through R&D
(research and development)
• We saw the “top dog” effect – firm 1
overinvests in R&D to win market share from
rival
A positive analysis: Strategic R&D
• Preceding analysis treats R&D as a routine
matter – a bit more R&D produces a bit more
cost reduction – and not all innovation is like
that
• Innovation might be in regard to production
processes for existing products….but might
introduce entirely new products
• Innovation might be incremental….but might
happen in occasional big steps
• Uncertainty is central to innovation –
innovation involves luck
Market structure & incentives for R&D
• Product or process innovation?
– process innovation easier to analyse, so start with
that
• Small or large innovation
– small is also more straightforward, so start there
• Usual terminology is drastic versus non-drastic
innovation
– a drastic innovation is one where the monopoly
price corresponding to the new, lower costs is
below the original level of marginal costs
• So we proceed with
– a process innovation
– ….that is non-drastic
– ….and that is non-stochastic (i.e. certain)
• First firm to invent process can obtain an
infinitely lived patent i.e. a property right that
prevents anyone else from using the invention
at any future time
• What is the social value of the innovation? i.e.
net increase in surplus if p=mc before and
after innovation (efficient pricing)
• The innovation reduces marginal cost from
high level, 𝑐 ℎ , to low level, 𝑐 𝑙 .

• Gain in consumer surplus illustrated in figure


18.2: label this gain 𝑉𝑆
• Compare this with an initially competitive
industry
• Pre-innovation, price equals 𝑐 ℎ
• Innovation means one firm becomes a
monopolist, protected by a patent
• Post-innovation, price equals 𝑐 ℎ (or, more
precisely, 𝑐 ℎ − 𝜀, since just undercuts rivals)
• Demand and profit are 𝐷 𝑐 ℎ and
𝑐 ℎ − 𝑐 𝑙 𝐷 𝑐 ℎ , which is dark-shaded area in
Fig 18.2. Label this 𝑉𝐶
• What about a pre-innovation monopoly?
• Sets monopoly price before, 𝑝𝑀 𝑐 ℎ
• …and monopoly price after, 𝑝𝑀 𝑐 𝑙
• But note that is not constrained by rivals, so
• 𝑝𝑀 𝑐 𝑙 > 𝑐 ℎ
• Value to innovator, 𝑉𝑀, can be shown to be
• Can see that 𝑉𝑀 < 𝑉𝐶

• So the value of innovation is less to an


established monopolist….than it is to a
competitive firm that becomes a monopolist

• Labelled the replacement effect

• By innovating, a monopolist is replacing itself,


and losing the profits it previously earned i.e.
the innovation consists of the replacement of
one profitable monopoly with another.
• Tempting to argue that monopolist has less
incentive to innovate
• But this neglects competition in innovation
(which preceding argument effectively
excluded)
• So think of existing monopolist competing
with a potential entrant to make the same
innovation as before
• The efficiency effect of Gilbert and Newbery
(1982) says monopolist will have greater
incentive to innovate
• To see this, suppose entrant makes the
innovation: its profits are 𝜋 𝐷 𝑐 𝑙 , 𝑐 ℎ

• In this case monopolist becomes a duopolist:


its profits are 𝜋 𝐷 𝑐 ℎ , 𝑐 𝑙 (incumbent firm
would become high-cost firm in a duopoly)

• By contrast, if monopolist innovates its profits


are 𝜋 𝑀 𝑐 𝑙
• Monopoly profits are higher than industry
profits in duopoly:
• 𝜋 𝑀 𝑐𝑙 > 𝜋 𝐷 𝑐ℎ , 𝑐 𝑙 + 𝜋 𝐷 𝑐𝑙 , 𝑐ℎ
• Rearranging
• 𝜋 𝑀 𝑐𝑙 − 𝜋 𝐷 𝑐ℎ , 𝑐 𝑙 > 𝜋 𝐷 𝑐𝑙 , 𝑐ℎ

• Left-hand side is monopolist’s gain from


winning rather than losing….right-hand side is
equivalent for entrant
• What does this show?
• If an innovation is bound to happen – and only
question is who makes the innovation – then
monopolist has greater incentive to innovate
• Why? Because entrant innovation reduces
industry profits and incumbent innovation
preserves them
• Corollary of this: a monopolist would always
be prepared to outbid entrant for an
innovation and thus preempt the entry
• A patent race model has the following
elements:
– rivals spend money on research until one of them
makes an innovation
– the more that is spent, the earlier the innovation
is expected to occur
– the innovation process is stochastic – involving
uncertainty – not deterministic
• By spending more on d (R&D) an innovator
can bring forward the time of innovation T

• But there are diminishing returns to this


investment
Time to
successful
development

Cost incurred
• Because of discounting, an earlier innovation
is more valuable than a later one
• If innovation was at end of year T would
discount back to present by dividing by
1 + 𝑟 𝑇 , where r is the interest rate
• For example:
PV=100/(1+0.1)5 =100/1.61= 62.11
PV=100/(1+0.1)10 =100/2.59=38.61
• To allow for within-year innovation treat time
as continuous, not discrete
• If innovation is at time T therefore discount by
multiplying by 𝑒−𝑟𝑇
• A monopoly innovator – without threat of
entry – would therefore choose d to maximise
• 𝑒−𝑟𝑇 𝑑 𝑉𝑀 − 𝑑
• Call the resulting R&D choice 𝑑𝑀
• Fig 18.6 graphs the two parts of the objective
and illustrates 𝑑𝑀
Time to
successful
development

Cost incurred
• On the diagram 𝑑𝑆 is the socially optimal
expenditure
• 𝑑𝑆 > 𝑑𝑀 because 𝑉𝑆 > 𝑉𝑀 : a product priced at
cost is worth more to society, and so its worth
spending more resources to get it
• The speed of research undertaken by a
monopolist is less than the socially optimal level
• The diagram also shows the competitive
equilibrium research 𝑑𝑒
• In a patent race, expenditure on R&D will be bid
up by the “race to be first” to the point at which
the winning firm will expect its revenues under
the patent to be exactly equal to the costs of
doing the research. The losing firms will spend
nothing and earn nothing.
• You can see that the level of R&D is excessive
in this case: 𝑑𝑒 > 𝑑𝑆
• Competing innovators will enter the race until
the expected rewards are reduced to cost
• Each innovator ignores the effect its decisions
have on the success probability of others
• Outcome is like the “tragedy of the commons”,
e.g. when an open access fishery is over-
fished from society’s point of view
• Adding uncertainty makes a patent race more
realistic
• No amount of expenditure will guarantee any
firm will win a patent race,
• A firm can increase its probability of winning by
increasing its expenditure
• The more firms ‘racing’ the earlier the likelier
date of innovation.
• We would expect to see more than one firm
committing R&D expenditure in equilibrium for a
given patent race.

You might also like