Upload
precious-thomas
View
20
Download
0
Tags:
Embed Size (px)
DESCRIPTION
Using Economic Theory for Policy E. Maskin Institute for Advanced Study. 25 th Anniversary of FEDEA Madrid November 10, 2010. Many economists (including me) attracted to economic theory because besides being intellectually fascinating thought it could make world a better place - PowerPoint PPT Presentation
Citation preview
Using Economic Theory for Policy
E. MaskinInstitute for Advanced Study
25th Anniversary of FEDEAMadrid
November 10, 2010
2
• Many economists (including me) attracted to economic theory because
– besides being intellectually fascinating
– thought it could make world a better place
• Will argue that, economic theory has made world better place
• Two distinct contributions
(1) positive
(2) normative
3
(1) Positive Economic Theory• economics helps us understand how particular economic
institutions work - - or don’t work
• take the market for navel oranges‒ if you’re in charge of getting navel oranges produced and allocated
to consumers, want to maximize
consumers’ welfare ‒ producers’ cost
• How would you do this?
let consumer 's welfare from consuming 1, ,i i iw x i x i n
producer 's cost of producing 1, ,j j jc y j y j m
max such that i i j j i ji j i j
w x c y x y
4
• One way–
–
–
• Another way– set price p
–
–
–
have each consumer report ii w
have each consumer report jj c
compute , y that solve maximizationi ix
adjust until i ji j
p x y
allow each consumer to demand what he wants, , at ix p
allow each producer to supply what it wants, , at jj y p
5
• second method is essentially competitive market mechanism– efficient - - solves maximization– requires transmitting a lot less information than first
method• one can show competitive mechanism
– requires less information transmission than any other efficient mechanism
Hayek (1945)Hurwicz (1977)Jordan (1982)
• In fact, can do away with centralized price setter– (unconstrained) bilateral bargaining leads to efficient
outcomeGale (1987)
6
• Thus, economic theory gives us great insight into when and how competitive markets work well: if (as in navel orange market)– many agents (consumers and producers)
– no significant externalities
then competitive markets generate efficient outcome
Arrow and Debreu (1954)
affects only i ix w
affects only j jy c
7
• Theory also shows when markets don’t work well• Example - - air pollution
‒ for optimum
‒ for equilibrium: p = price of pollution reduction
‒
• so, air pollution not reduced efficiently through market
benefit to customer of reductions i k kk
w x i x
cost to producer of reducing pollution by j j jc y j y
max s.t.i k j j k ji k j
w x c y x y
first-order condition for all i k j j
i k
w x c y j
optimum equilibrium Samuelson (1954)
i k j jk
w x p c y
8
Today, more interested in:
(2) Normative Economic Theory• how can economic theory be used to create
better institutions?‒ to formulate good policy?
9
2 Examples• drawn from issues I’ve worked on myself• in increasing order of theoretical and empirical
difficulty
10
Example 1 – pollution reduction• argued before that competitive market cannot
adequately deal with pollution• But major contemporary problem ‒global
warming‒ is pollution problem (greenhouse gas emissions)
• What to do?
11
• several years ago, British government wanted to induce major greenhouse gas-emitting producers to reduce emissions voluntarily
• budget = $500,000,000• Challenge: allocate budget in way that
maximizes emission reduction
12
More precisely,• budget B
• for each firm j, j = 1, … , m
–
– good approximation:
• mechanism:
–
– outcome : each firm j assigned
firm 's cost of reducing by unitsj j jc y j y
, 1
, 1
j j j
j j
j
c y yc y
y
ˆeach firm bids jj c
ˆreduction jyˆtransfer jb
13
• Transfers must satisfy
•
•
subject to
budget constraint
individual rationality
incentive compatibility
ˆ jb Bˆ ˆ 0 j j jb c y
ˆ in equilibriumj jc c
ˆProbelm: maximize E jy
individual rationality
budget constraint
incentive compatability
14
Economic theory gives us:
Solution: (assuming low costs more likely than high costs)• if no bids less than B,
• if, for some k, exactly k bids less than B/k
• if, for some k, more than k bids less than B/k but fewer than k + 1 less than B/(k + 1)
Generalizes to different capacities for different firms
1ˆ ˆ 0my y
ˆ1, /ˆ
ˆ0, /
j
j
j
c B ky
c B k
1 1ˆ ˆ,ˆ0, otherwise
k kj
j
c c cb
ˆ/ , /ˆ
ˆ0, /
j
j
j
B k c B kb
c B k
st1ˆ1, 1 lowest bidˆ
0, otherwise
kj
j
c c ky
15
Advantages for economic analyst:
•
•
• government controls mechanism
‒ we (analyst) know what game is
‒ firms know what game is
• no extensive empirical work needed
‒
ˆclear goal maximize iE y limited uncertainty only about jc
except for distribution of jc
16
Example 2 ‒ Software Patents
Should software be patentable?• policy issue still open in EU• To answer question
‒ need theoretical framework that incorporates both costs and benefits of patenting
‒ need to formulate policy objective
‒ then asses parameter values empirically
17
Benefit: patents encourage innovation• inventor cannot make discovery without
incurring costs• without patent, discovery can be freely imitated• imitation reduces inventor’s revenue from
discovery• so, inventor may not even cover costs• anticipating this, inventor may decide against
incurring costs in first place• society deprived of valuable innovation
18
Costs
(1) patents confer monopoly power• but this argument by itself doesn’t outweigh benefit
• can adjust patent horizon
• many significant innovations attributable to patents
e.g., most modern drugs‒ given huge R&D costs, need high-powered incentive schemes
patents or prizes
‒ advantage of patents over prizes: don’t have to know goal in advance
19
(2) patents may interfere with innovation• consider
‒ software‒ computers‒ semiconductors
• all highly innovative industries, but‒ weak patent protection (at least in early history)‒ frequent imitation
• Natural experiment in U.S. with software‒ before 1985 software not protected‒ after 1985, courts ensure enforceability‒ but decrease in software R&D per firm‒ similar story in Japan
20
What’s different about software (as compared to drugs)?• innovation highly sequential• rather than single big break-through
‒ each innovation is small and builds on previous innovations
21
Why does sequentiality make big difference?• suppose I’ve made a discovery (mousetrap) and
put a patent on it• suppose you have idea how to build on my
discovery (better mousetrap) • because you’re using what I’ve discovered, my
patent can block you• so, then my patent reduces innovation
22
Not complete argument• if your follow-up R&D so important• then should be profitable• but then why don’t I license discovery to you?• could presumably capture some profit by
charging license fee
23
Problem: what fee will I set?• I am a monopolist
‒ have exclusive right to my discovery
• economic theory indicates that monopolist has incentive to set prices too high
• So, I am likely to set license fee so high that‒ significant risk you won’t pay it
‒ so, significant risk your R&D will be blocked
24
• So patents can interfere with follow-on innovation• What about original discoveries?
‒ aren’t patents still needed to induce me to invent mousetrap in first place?
• Answer: not as much as in case of non-sequential innovation‒ with sequential innovation, if I decide to incur R&D costs,
raise probability of mouse trap discoveryalso raise probability of follow-on discoveries
‒ so potential benefit of R&D bigger‒ if I capture just fraction of benefit without patents, then patents not
so important
25
Thus, in industry with sequential innovation• patents reduce follow-on innovation• although patents still promote original
innovation,‒ not so important as in nonsequential world
• thus, if industry is sequential enough‒ we may be better off without patents
26
• This is theoretical argument
• Forms framework for empirical test‒ Bessen and Hunt (2004)‒ suggests that in U.S. software patents have been
detrimental to welfare
27
• Use of normative economics more difficult than in carbon mechanism case– don’t know exact game that software inventors are playing– game much less under analyst’s control– many parameters uncertain - - cost of R&D, value of discovery,
degree of sequentiality,...– What is right policy objective? (max consumer + producer
surplus?)
• But theory perhaps even more important in patent case‒ there are so many factors that could be important‒ theory enables us to focus on a few that are truly important, e.g.,
sequentiality
• Economic theory not only useful‒ good policy would be hopeless without it