Transcript A “Drastic” Innovation

Research and Development
Part 1: Innovations and Patents



Economic growth is caused
primarily by technological
progress (the Solow model of
growth).
R&D is the “engine of
technological change”.
Firms become industry leaders
by conducting R&D that leads to
innovations in products and
processes.


Schumpeter coined the phrase
“creative destruction” to refer
to the capitalist system.
Periodically, new products and
processes are developed which
destroy the market power of old
products and processes.
What is R&D?



Basic Research: Does not lead directly
to new product or process, but
improves “fundamental knowledge”.
Applied Research: Involves substantial
engineering input and results in new
product or process.
Development: Move product/process
to consumer market/mass production.
What is R&D, con’t




Process innovations: better method
for producing existing product.
Product innovations: new product.
Drastic innovation: dramatically
reduce costs to extent that innovator
essentially becomes a monopolist.
Non-drastic or gradual innovation:
improves firm’s competitive position,
but there still is competition.
A “Drastic” Innovation
$
PM
PC
MC for
industry
MR
QC
QM
D
New MC
Quantity
A “Non-Drastic” Innovation
$
PM
PC
MC for
industry
New MC
MR
QM
QC
D
Quantity
R&D and Market
Structure


What market structures are most
conducive to R&D?
Early View (Schumpeterian Hypothesis):
 R&D
generally conducted by large firms.
 R&D generally conducted in industries with
market power.
 Thus, markets with large firms and market
power not necessarily bad as this would
encourage R&D and innovation.
Market Structure and
the Incentive to Innovate



Is the Schumpeterian Hypothesis
correct? Are incentives for R&D affected
by market structure?
Start with a basic inverse demand curve:
P= A-BQ.
We want to consider an innovation that
could increase social welfare and then see
whether market structure affects
whether such an innovation will occur.
Arrow Model of Mkt.
Structure & Innovation



Assume that the innovation would cost K
to research and develop.
The result of the innovation is an
improvement in the production process so
that the constant marginal cost of
production drops from c0 to c1.
Compare expenditure, K, to potential
increase in surplus from change in MC,
(appropriately discounted to account for
increased surplus in the future as well).
The Welfare Effects of an Innovation
$
CS
C0
C1
D
Quantity
Arrow Model, con’t


In a competitive market, a firm will only
innovate if the cost of innovation is less
than the increased profit as a result of
the innovation.
Consider a perfectly competitive market.
 If
the innovation could be adopted by
other competitors, there would be no
increase in profit. All firms would adopt,
but because of perfect competition, price
would drop to equal the new lower cost.
Arrow Model, con’t


If the market is perfectly competitive
but the innovation is not adoptable, the
firm will be able to make profit as a result
of the innovation.
The firm can undercut the competitive
price by a couple of cents and supply
essentially the entire market.
Increase in Profit in a Competitive Market
from a Non-Adoptable Innovation
$
Innovator’s
price
Note that the increase in profit
is less than the increase in
potential welfare
Increase
in Profit
C0
C1
D
Quantity
Increase in Profit in a Monopoly from
Innovation
$
Note that the increase in profit for the
monopolist (the area of the blue box
minus the area of the yellow box) is less
than the increase in profit for the firm
in a competitive market.
Innovator’s
price
C0
C1
MR
D
Quantity
Arrow Model, con’t

A monopolist will actually value the
innovation less than a competitive firm.
 The
monopolist was already making profit,
already had some market power. This is
known as the replacement effect.


Even with more complex models, general
result is that the more competitive the
market, the more gains to an innovator.
In all cases, firms undervalue innovation
compared to effect on total surplus.
Arrow Model, con’t




Is the Schumpeterian Hypothesis wrong?
Are gains from innovation inversely related
to market power?
Market power pre-innovation makes the
post-innovation gain smaller.
But we assumed that post-innovation the
market would be a monopoly, regardless of
the pre-innovation structure.
Also, we said nothing about the market for
innovation itself.
Another Model of Mkt.
Structure & Innovation



Suppose that there are two potential
innovators, an incumbent monopolist with
technology c0 and a potential entrant.
Assume the innovation is non-adoptable, so
that the first firm to innovate can lower
marginal costs to c1 at a cost of K.
If the potential entrant innovates first, he
becomes the low-cost firm in a duopoly. His
return to innovation is D(c1,c0) - K.
Another Model of Mkt.
Structure & Innovation




If the monopolist is the first innovator, he
will retain his monopoly and earn profits of
M(c1) - K.
If the monopolist does not innovate, the PE
will, so the monopolist will get D(c0,c1).
Thus the return to innovation for the
monopolist is M(c1)-D(c0,c1) - K.
For the PE, the return is D(c1,c0) - K.
Another Model of Mkt.
Structure & Innovation


As long as M(c1)-D(c0,c1) - K > D(c1,c0) - K
the return to innovation is greater for the
monopolist than for the PE.
Rewrite this condition:
M(c1) > D(c1,c0) + D(c0,c1).


Which says that if the profit to a
monopolist is greater than the joint profit
of duopolists, the gain to innovation is
greater for a monopolist. And it is.
This is known as the efficiency effect.
Schumpeter Revisited




Is the Schumpeterian Hypothesis wrong?
It depends on the nature of competition.
If the firm starts with market power, the
increase in market power due to an
innovation will not be as large as if the firm
started without market power.
However, firms with market power have
greater incentives to innovate to protect
their existing market power.
Even More Evidence: the
Dasgupta/Stiglitz Model





Consider a slightly more sophisticated model
where all firms in the industry can innovate
simultaneously.
Assume firms compete through quantity.
i = P(Q)qi - c(xi)qi - xi where x is the firms
expenditure on R&D.
If all firms spend the same amount, x*, on
R&D, all firms have the same cost c(x*).
Solve just as we would any n-firm Cournot
model.
Dasgupta/Stiglitz con’t



We then get the following expression of the
equilibrium quantity for each firm in this
market:
q* = (a-c(x*))/(b(n+1)).
To determine the optimal level of R&D, take
the derivative of profit w.r.t. xi and set
equal to 0.
i = P(Q)qi - c(xi)qi*- xi , so the derivative is:
-(dc(xi)/dxi)qi* - 1 = 0.
Dasgupta/Stiglitz con’t



What does -(dc(xi)/dxi)qi* - 1 = 0 mean?
Remember that dc(xi)/dxi is the decrease in
marginal cost due to an additional dollar
spent on R&D. So the first term is the
total benefit of an additional dollar spent on
R&D.
In equilibrium, this total benefit must be
equal to the cost of the additional R&D, i.e.
$1.
Dasgupta/Stiglitz con’t




But the important thing to note is that the
marginal benefit -(dc(xi)/dxi)qi* , depends
on q*.
And q* depends on the number of firms in
the industry: q* = (a-c(x*))/(b(n+1)).
The more firms, the lower is q* and thus
the lower the benefit from R&D.
A lower benefit means less R&D spending, so
the more firms in the industry, the less
spending by each firm on R&D.
Dasgupta/Stiglitz con’t



How will total industry spending, nx*, be
affected by concentration?
Total spending may increase or decrease as
industry size increases, depending on the
elasticity of demand in the industry.
However, in most cases, increasing the
number of firms leads to less industry
spending.
R&D and Market Power


How do we balance the potential positive
aspects of market power (w.r.t. R&D and
innovation) with the positive aspects of
less concentrated markets (w.r.t.
efficiency)?
Key is to recognize the dynamic aspect of
the process. We want to increase
efficiency over time which may require a
settling for less efficiency in the short
run.
Public Policy w.r.t. R&D


What policies best encourage the
discovery of new products and processes?
What policies best encourage
dissemination of new ideas and
technologies?
Patents and Copyright



Patents and copyright confer property
rights to new inventions, new designs, and
new creative works.
The property rights allow the innovator to
exert monopoly power, which acts as an
incentive to encourage R&D and innovation.
However, to achieve the efficiency gains
from the innovation, it must be used
widely, so the property rights must be
terminated at some point.
Effect of Patent on Surplus
This surplus goes to consumers (A)
$
This surplus goes to producer
over length of patent, then
transfers to consumers (B)
This surplus goes to consumers
when patent expires (C)
MR
QP
D
QC
Quantity
What is the optimal
patent length?




If T is the length of the patent, then the
producer gets B (from the graph) for T
years-appropriately discounted of course.
The consumers get the A each year, plus
the B and C for all years after T.
Note that after T years, B just transfers
from the producer to the consumer.
Real gain in surplus is C, which we get
after the patent expires.
What is the optimal
patent length, con’t




The shorter T, the sooner we get C.
However, the producer decides whether
to innovate based on the increased profit,
i.e. how long he gets the surplus in B.
The longer T, the more incentive to
innovate and the bigger are A, B, and C.
Optimal patent length balances these two
effects.
Optimal Breadth of
Patents



Breadth is the amount by which innovation
must differ from an existing product or
process.
The more broad a patent is, the harder it
is to “invent around” the patent and thus
the more profit the producer can get.
Related to optimal length -- i.e. could have
“short and fat” patents or “long and thin”.