H5: Innovation and intellectual property
Introduction
- So far looked at technology as given
- Discussed H/V differentiation, assuming firms select it without risk
- Improving quality or creating new products is expensive and uncertain
- How innovation affects firms competition, and how strategic relations affect
the aggregate innovative output
Innovation in economics
- Expansion of the production set
- New input-output combination that was unavailable before
=> New output with existing inputs (new products)
=> Known output with new combination (More efficient processes
- “New” product is rare, is in reality like old product with higher value
=> Still can be a substitute to existing products
=> Large enough cost reduction can create a new market
- Classify innovation based on effect they have on firms: drastic / non drastic
Non-drastic (Incremental) vs drastic innovation
=> Drastic innovation is such that, provided that no can imitate, can be a
monopolist
=> Non-drastic innovation is such that the monopoly price is higher than the
marginal cost they would set before: limit pricing equilibrium, set price so others
can’t undercut the price
=> Larger profits in case of drastic innovation
What market structure is the most innovative
=> 2 Main (contrasting) theories
- Schumpeter: Innovation manifest as “creative destruction”, monopolists are
vulnerable to new entrants innovation and they can use their profits to shield
themselves from this risk by innovating first
- Arrow: A monopolist has more to lose from changing the status quo.
Conversely, competitive firms can only benefit from escaping the competition
1
,Arrow model
- Assume linear demand, constant marginal cost c0
- Firms can pay k(η)v = kη2 to obtain a drastic innovations with probability
η and reduce marginal cost by ∆c
- Arrow compares the decision of a monopolist and a competitive firm
- π0M= monopolist market profits without innovation
- πID = duopolist market profits if innovation is successful
- π0D= duopolist market profits without innovation
=> With Bertrand competition: π0D = 0
- Bertrand firms: πC(η) = ηπIM - k(η)
- Monopolist: πM(η) = ηπIM + (1 - η)πM - k(η)
=> Replacement effect (1 - η)πM: Difference new profit and old profit,
cannibalization of profit old product
=> Extra term: if they fail, they still make a profit
- Firms facing competition have strictly higher incentive to invest
- Not because they gain more, but they only gain in the scenario of innovation
Schumpeter idea
- Monopoly position can be lost because of innovation of other firm
- Need extensive game to capture this scenario
- Two players: Incumbent and potential entrant
- I has MC = c and makes profit π0M, E currently out of the market cE >> c
- Firms can invest K(η) = η2/2 to reduce their MC to c’ < c << cE with
probability ηi ∈ [0,1]
- Two-stage game:
- Incumbent chooses its investment ηI first
- Entrant observes I’s choice and chooses ηE
- Assume innovation as non drastic to challenge idea of Schumpeter
Schumpeter Payoffs
- If neither succeeds: profits stay the same π0M for I and 0 for E
- If bot succeed: Betrand competition with equal MC = c’ => zero profits
- If I succeeds and E doesn’t: I profits increase to π1M > π0M
- if E succeeds and I doesn’t: limit-pricing equilibrium (non-drastic
innovation). E makes profit πE < π0M and I makes 0
- Payoff normalization: K’(η=1) = 1>π1M>π0M>πE
=> Guarantees that investing the maximum η=1 is not always optimal
2
, => If incumbent invests more and are more likely to succeed, Entrant can’t win
=> If I doesn’t invest and E does, best case scenario for I is to get 0 profit
=> Model includes both replacement effect and Shumpeter’s idea
Illustrative example sl17
- Key idea: The incumbent internalizes the effect of its investment on the
entrant's decision and a result invests more to preempts the entrant’s
innovation
- Key assumptions:
- Non-radical innovation
- The monopolist has more to lose than the entrant has to gain (π 0M>πE)
Arrow vs Schumpeter: conclusion
- Both the Arrow replacement and preemptive effect exist and documented
- The prevailing force depends on the industry and the type of competition
(e.g. the magnitude of the profits we set)
- Aghion et al. (2005) propose a richer model, where many firms can innovate
on several product lines with different characteristics. Find non-linear effect
of competition
- Products in “too competitive" markets have too little pro˛ts to be
appealing (same as low vE )
- Products with “too little" competition, Arrow replacement dominates,
under-investment by
- They calibrate the model on US granted citation-weighted patents
(innovation) and Lerner index/markups (competition)
- Find an inverted-U relation (their relation isn’t linear!)
- Study is a bit flawed, but we still believe in this relation
3
Introduction
- So far looked at technology as given
- Discussed H/V differentiation, assuming firms select it without risk
- Improving quality or creating new products is expensive and uncertain
- How innovation affects firms competition, and how strategic relations affect
the aggregate innovative output
Innovation in economics
- Expansion of the production set
- New input-output combination that was unavailable before
=> New output with existing inputs (new products)
=> Known output with new combination (More efficient processes
- “New” product is rare, is in reality like old product with higher value
=> Still can be a substitute to existing products
=> Large enough cost reduction can create a new market
- Classify innovation based on effect they have on firms: drastic / non drastic
Non-drastic (Incremental) vs drastic innovation
=> Drastic innovation is such that, provided that no can imitate, can be a
monopolist
=> Non-drastic innovation is such that the monopoly price is higher than the
marginal cost they would set before: limit pricing equilibrium, set price so others
can’t undercut the price
=> Larger profits in case of drastic innovation
What market structure is the most innovative
=> 2 Main (contrasting) theories
- Schumpeter: Innovation manifest as “creative destruction”, monopolists are
vulnerable to new entrants innovation and they can use their profits to shield
themselves from this risk by innovating first
- Arrow: A monopolist has more to lose from changing the status quo.
Conversely, competitive firms can only benefit from escaping the competition
1
,Arrow model
- Assume linear demand, constant marginal cost c0
- Firms can pay k(η)v = kη2 to obtain a drastic innovations with probability
η and reduce marginal cost by ∆c
- Arrow compares the decision of a monopolist and a competitive firm
- π0M= monopolist market profits without innovation
- πID = duopolist market profits if innovation is successful
- π0D= duopolist market profits without innovation
=> With Bertrand competition: π0D = 0
- Bertrand firms: πC(η) = ηπIM - k(η)
- Monopolist: πM(η) = ηπIM + (1 - η)πM - k(η)
=> Replacement effect (1 - η)πM: Difference new profit and old profit,
cannibalization of profit old product
=> Extra term: if they fail, they still make a profit
- Firms facing competition have strictly higher incentive to invest
- Not because they gain more, but they only gain in the scenario of innovation
Schumpeter idea
- Monopoly position can be lost because of innovation of other firm
- Need extensive game to capture this scenario
- Two players: Incumbent and potential entrant
- I has MC = c and makes profit π0M, E currently out of the market cE >> c
- Firms can invest K(η) = η2/2 to reduce their MC to c’ < c << cE with
probability ηi ∈ [0,1]
- Two-stage game:
- Incumbent chooses its investment ηI first
- Entrant observes I’s choice and chooses ηE
- Assume innovation as non drastic to challenge idea of Schumpeter
Schumpeter Payoffs
- If neither succeeds: profits stay the same π0M for I and 0 for E
- If bot succeed: Betrand competition with equal MC = c’ => zero profits
- If I succeeds and E doesn’t: I profits increase to π1M > π0M
- if E succeeds and I doesn’t: limit-pricing equilibrium (non-drastic
innovation). E makes profit πE < π0M and I makes 0
- Payoff normalization: K’(η=1) = 1>π1M>π0M>πE
=> Guarantees that investing the maximum η=1 is not always optimal
2
, => If incumbent invests more and are more likely to succeed, Entrant can’t win
=> If I doesn’t invest and E does, best case scenario for I is to get 0 profit
=> Model includes both replacement effect and Shumpeter’s idea
Illustrative example sl17
- Key idea: The incumbent internalizes the effect of its investment on the
entrant's decision and a result invests more to preempts the entrant’s
innovation
- Key assumptions:
- Non-radical innovation
- The monopolist has more to lose than the entrant has to gain (π 0M>πE)
Arrow vs Schumpeter: conclusion
- Both the Arrow replacement and preemptive effect exist and documented
- The prevailing force depends on the industry and the type of competition
(e.g. the magnitude of the profits we set)
- Aghion et al. (2005) propose a richer model, where many firms can innovate
on several product lines with different characteristics. Find non-linear effect
of competition
- Products in “too competitive" markets have too little pro˛ts to be
appealing (same as low vE )
- Products with “too little" competition, Arrow replacement dominates,
under-investment by
- They calibrate the model on US granted citation-weighted patents
(innovation) and Lerner index/markups (competition)
- Find an inverted-U relation (their relation isn’t linear!)
- Study is a bit flawed, but we still believe in this relation
3
