1. The document discusses the economics of patents and how they can be designed to balance incentives for innovation with allowing useful spillovers. 2. It analyzes optimal patent length, breadth, and height using economic models. Longer patents create more profits but less consumer welfare. Broader patents sacrifice consumer surplus for small profits. 3. Patents should be designed prudently to avoid locking out relevant implications for others and leave room for spillovers, as knowledge has a systems nature and inventions build on each other.

1. The economics of patents
Merit course – 2005
Motivation: to see “how patents work” and
how economic theory can help to design
them
the role of patents
patent length
Other patent dimensions

2. IBM vs Apple
Winners & losers?
1000
45
40
Microsoft
35 100
30
25
Intel
10
20 Compaq
15 Apple
10 1
5 IBM
0
0
1981
1983
1985
1987
1989
1991
1993
1995
1997
1985 1987 1989 1991 1993 1995 1997
Bron: Harvard Business School Apple case studies 1992 & 1998 Bron: Worldscope database

3. Technology Spillovers and incentives
The problem that spillovers create is an
incentive problem
This may be solved by the award of a patent:
the legal monopoly to be the only user of an
invention
Monopolies make an economist alert!
Technology spillovers are not totally bad: they
are a source of welfare
– Example: the history of the personal computer

4. The role of patents
The three Ps:
– Patronage (e.g., universities)
– Procurement (e.g., defense or public sector
research institutes)
– Patents
Patents and spillovers: publication of
patented inventions

5. Designing patents
Can we design a patent that strikes the
balance between incentive and welfare
generation?
Patent length, breadth, width and height

6. Patents and spillovers – How can
patents leave spillovers?
Patentable and non-patentable parts of
knowledge
Standing on the shoulders of giants
Rent spillovers

7. Nordhaus model of patent length
Minor and drastic innovations

8. Nordhaus model of patent length
Longer patents create a longer profit stream
for the inventor, but at a decreasing rate
(profits far away in the future are worth less)

9. Nordhaus model of patent length
Maximum profit for a monopolist
minor
(A Bc) (c c),
qx
A
Value of the consumer surplus
S cd.
B
0

10. Nordhaus model of patent length
Invention possibilities function
cc
R,
c
T
NPV of net profits
minor
V e d R.
0
T T
1e
V c R (A Bc)e d R c R (A Bc) R.
0
T
V 1e
(1 )
cR (A Bc) 10
R
Optimal R&D
1
T
1e
expenditures
1
R c(A Bc) .
T T
1e 1e
1
V c c(A Bc) (A Bc)
Resulting NPV of
1
profits
T
1e 1
c(A Bc) .

11. Nordhaus model of patent length
NPV of profits for varying patent duration

12. Nordhaus model of patent length
Longer patents create less consumer welfare
(they leave a shorter period for consumers to
benefit from all the advantages of the patent)

13. Nordhaus model of patent length
Total welfare resulting
T
e
WV e d V .
from innovation
T
q x(c)
A 1
Consumer surplus after
B(c c)2,
cd
B 2
patent runs out
q x(c)
2
Part of total welfare due
T T T
e 12 1e e
2 1
Bc c(A Bc)
2
to increased consumer
T T
1e e
1
c (A Bc) c(A Bc) .
surplus

14. Nordhaus model of patent length
Consumer surplus (NPV) as a function of
patent duration

15. Nordhaus model of patent length
Total welfare as a function of patent duration
Higher demand elasticity decreases optimal patent life, higher
technological opportunities leads to shorter optimal patent life

16. Patent breadth
Systems nature of knowledge: spillovers are
often in the form of relevant implications by
outsiders
In the spirit of the Nordhaus model: be careful
not to prevent too many spillovers from
locking out
Modeling patent breadth: Horizontal product
differentiation

17. Van Dijk / Klemperer model of patent
breadth – Horizontal differentiation
vp d if the consumer buys
U
0 otherwise
b p The indifferent
vp |w 0| v |w b| w .
2 2
consumer

18. Van Dijk / Klemperer model of patent
breadth
1
pb p
Profit function
.
2 2
1
b ( 1)p b
0 p ,
Find optimal price
p 2 2 1
1
Substitute into indifferent
b b b
w , pw .
2( 1) 2( 1)
consumer
b
Consumer welfare loss “right”
b2 ( 2)2
(b )d .
1
1)2
8(
w
w
b 2 (4 )
Consumer welfare loss “left”
(p )d .
2
8 ( 1)
0
1
b2 ( 2)2 b 2 (4 ) b b
.
Total welfare loss
1 2
2( 1) 1
1)2 8 ( 1)2
8(

19. Van Dijk / Klemperer model of patent
breadth
1 1
2)2
b( b b (4 ) b
1 2
1.
4( 1) 1 4( 1) 1
Broad patents are bad,
because they sacrifice a lot of
consumer surplus for little
profits (incentives)
Narrow patents are bad,
because they sacrifice a lot of
consumer surplus for little
profits (incentives)
Patent breadth is a tool to vary the ratio of welfare loss to
incentive (=profit). Policy: pick the best ratio and adjust
patent length

20. Patent breadth
‘Gold mining’ of patents
– Genetic technology
– The importance of applicability of patents
A trend towards broader patents (pro-patent
era in the US)
Software protection: copyright (narrow) vs
broad patents: EU parliament debate

21. Prudence in patent breadth
Systems nature of knowledge: one invention
depends on another one
A prudent rule is to leave an important part of
the relevant implications to outsiders (do not
award broad patents on all imaginable
applications of a discovery)
Genetic research is an important
contemporary example

22. The last aspect of patent protection:
Inventive step
Every patent must have a minimum degree of
novelty and must not be obvious to “someone
skilled in the art”
– But check some of the patents in patent databases
on the web
– Would you consider the “one click to buy” system
at Amazon.com “not obvious to someone skilled in
the art”?

23. Van Dijk model of patent height -
Vertical differentiation (quality)
v m f p if the consumer buys
U
0 otherwise
p2 p1
Indifferent consumer
m f1 p 1 m f2 p2 m ,
f2 f1
p2 p1 p2 p1
x1 , x2 1 .
Demand functions
f2 f1 f2 f1
p2 p1 p2 p1
Profit functions
p1 , p2 p2 .
1 2
f2 f1 f2 f1
p2 2p1 1
1
0 p1 p,
22
p1 f2 f1
p1 2p2
Optimal prices yield a game
1
2
1 0 p2 (p f f ).,
2121
p2 f2 f1
(Nash equilibrium)
f2 f1 f2 f 1
p1 , p2 2 .
3 3

24. Van Dijk model of patent height

25. Van Dijk model of patent height
Innovation possibilities function
f 2.
R(f )
f2 f1 f2 f1
2 2
Net profit functions
V1 f1 , V2 4 f2 ,
9 9
V1 1
2 f1 0 f1 0,
f1 9
Optimal inventive
V2 4 2
2 f2 0 f2 .
step
f2 9 9
When h 2/9 patent height does not become restrictive
When h>4/9 the imitator decides not to enter the market
When 2/9 < h < 4/9 the first-best choice of the imitating firm
is ruled out

26. Strategic patents
Patents are intended as legal protection from
imitation
The Yale survey: only one third of innovations
is protected by a patent
Why not the other two thirds?
novelty requirements
secrecy better alternative
afraid of inventing around
Can patents serve a different function than
just protection?

27. Patent strategies – Granstrand

28. Strategic patents
Costs of litigation are an important aspect of
strategic patenting
Strategic patenting puts more emphasis on
the non-competitive aspects of patents, and
hence on the inefficiencies rather than the
welfare-creating effects