Open Source:

1
Economics of Innovation Econ 124/PP190-4/PP290-4
Open Source A New (and Different) Way to
Organize Innovation
Stephen M. Maurer Goldman School of Public Policy
2
Overview
Introduction History Definitions Incentives S
tability What is the Role of Licenses
Leaders? Proprietary Code Coexistence or
Conflict? Limits to Open Source? Extensions
3
Introduction
The Patent Paradigm Edisons Idea Factory The
Open Source Enigma Linux, Apache, etc. What
Makes People Do It? Weird Explanations
The Gift Economy, Craftsmanship and The
Flow State.
4
History Definitions
History
Antecedents Folklore (e.g., The Odyssey) User
Innovation (Machine Tools, Sporting
Gear) Defensive Publishing Using Patents to
Regulate Technologies and Keep Them
Public Patenting Fission (Fermi, Szilard),
Deuterium (Urey) Plutonium (Seaborg).
5
History Definitions
History
Immediate Origins Academic Incentives Reputati
on Signaling Own-Use Self-Satisfaction Ide
ology Efficiency (Sharing Information) ATT/UN
IX Appropriability Complements
6
History Definitions
Definitions
Definitions Right to See Source Code Right
to Copy Right to Modify Use Richard
Stallman (Free Software) Is Open Source a
Moral Principle? cf. Moral Rights The Viral
Principle General Public License (GPL) Open
Software Berkeley UNIX License Lesser GPL
7
Incentives
Review
The Incentives Approach No Dominant Incentive
System Each Mechanism has Generic Strengths and
Weaknesses
8
Incentives
Review
Investment Based on Price Signal
Avoids Moral Risk Sponsors
Avoids Moral Risk Researchers
Avoids Undersupply
Elicits Private Information
v
v
v
v
Patents
Contract
v
v
Research
?
?
?
?
?
Open
Source
9
Incentives
Review
The (v c) Picture Welfare (v c)society
0 But, Private Utility (v - c)indiv.
0
10
Incentives
General
Many Different OSS Incentives Own-Use Signalin
g Education Complements Strategic
Goals Social Psychology Different Incentives
Welfare Implications for Each!
11
Incentives
Own Use
Incentive No. 1 Own-Use by Corporations and
Hobbyists
Investment Based on Price Signal
Avoids Moral Risk Sponsors
Avoids Moral Risk Researchers
Elicits Private Information
Avoids Undersupply
?
?
?
?
?
Open
Source
12
Incentives
Own Use
Investment Based on Price Signal First Answer
OSS Delivers (v c)indiv Patents Deliver (?
v c)society.
.
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
?
?
?
?
Open
Source
13
Incentives
Own Use
Investment Based on Price Signal Comments Cost
may be smaller than even one individuals
(v). OSS takes advantage of individuals with
unusual (v), (c). Own-Use by Oligopolists helps
aggregate demand (v). Explains bias toward
developer products. Doing some OSS projects is
better than doing none.
14
Incentives
Own Use
Second Answer Game Theory Monolithic
Software Delay Duplicated Work (v c) for
entire project Modular Software No
Delay Duplicated Work (v c) for every
module Individualized Software Jukebox
Model Economies of Scale vs.Competitive
Advantage
15
Incentives
Own Use
Elicits Private Information if (v c)indiv
0 (Open Source) (? v c)society
0 (Patents) But Coordination
Problems in both cases!
.
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
?
?
?
v
Open
Source
16
Incentives
Own Use
Agency Problems Researcher Sponsor! Cf.
Contracts Reporting Bugs, Programmer Skill
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
?
?
v
v
Open
Source
17
Incentives
Own Use
Avoids Undersupply No DWL!
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
v
v
v
v
v
Open
Source
18
Incentives
Signaling
Signaling Benefits For Individuals. Jobs,
Stock, Venture Capital Offers For Companies
Signaling Quality Assumes OSS Can Measure
Programmer Talent Better Open
Code Responsibility Self-Management Reusable
Code enhances human capital
19
Incentives
Signaling
Investment Based on Price Signal
Demonstrating Software Ability Is Only Weakly
Connected to Consumer Need. Skews products
toward developers system administrators No Free
Rider Problems.
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
?
?
?
Open
Source
20
Incentives
Signaling
Elicits Private Information?
Demonstrating Software Ability Is Only Weakly
Connected to Providing Information.
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
?
?
Open
Source
21
Incentives
Signaling
Agency Problems Incentive to Steal Credit The
Threshold Effect Is OSS More Transparent than
Corporate Life? Do Strong Programmers Self-Select?
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
?
v
Open
Source
22
Incentives
Signaling
Avoids Undersupply
No DWL!
Avoids Undersupply
Avoids Moral Risk Sponsors
Investment Based on Price Signal
Avoids Moral Risk Researchers
Elicits Private Information
v
v
Open
Source
23
Incentives
Education
Similar to Signaling But No Agency
Problems Does OSS Peer Review Facilitate
Education?
24
Incentives
Complements
Selling Complements Complements Can Include
Hardware, Applications, Marketing, Customization
Services, Education Tech Support Monopolists
Complete Appropriability Low-Cost Complements
Boost Revenue For Non-Monopolists Limited
Appropriability
25
Incentives
Complements
Implements a Pseudo-Patent System OSS
Delivers (? v c)society Patents Deliver (?
v c)society
.
.
Embedded LINUX Partial appropriability, more
spillovers than normal patent incentives.
26
Incentives
Strategic
Goals
Standard Setting Influencing OSS
Evolution Market Penetration Alumni
Effect Reassuring Consumers Hold-up Obsolesc
ence Switching Costs Preserving Option Value of
OSS
27
Incentives
Enhancements
Reintroduces Patent Incentives for Enhanced
Products Not GPL! When Would Originator Prefer
Berkeley UNIX? Welfare Implications Increases
Consumer Choice But What if Original Software
is Not Supported?
28
Incentives
Social
Psychology
Social Psychology Explanations Obligation,
Community Identification, Reputation, Ideology,
Altruism. Mechanics Re-label c as net
cost. Agency Problems Voluntarism
Markets Red Cross Blood Donations
29
Incentives
Social
Psychology
Fragility Fear of Exploitation Transparency and
Procedural Fairness Incentives are Weak But
Altruism in Games Small Rewards Can be
Important
30
Fixing Market Failures
1) IP Doesnt Work Perfectly Reliance on
Compiled Code Bug reporting, re-use, ability
to judge quality. Other Secrecy Costs Licensing
and enforcement costs. 2) Determining Programmer
Quality Closed Code Makes It Hard to Judge
Quality Peer Review Makes Process More
Transparent OSS offers a more accurate wage Can
Commercial Firms Use OSS Rankings? 3)
Determining Program Quality
31
Fixing Market Failures
4) Determining User Needs Inherent Difficulty
of Communicating Software Needs OSS and Price
Discrimination
32
Stability
Why Licenses? Welfare and Consumer
Sovereignty Possible Explanations Protects
Signaling/Reputation Incentives Berkeley UNIX,
not GPL Social Psychology Incentives and the
Non-Distribution Constraint Liability Discour
aging Defectors
33
Stability
Why Leadership? Certifying Quality Forking,
Hijacking and Trust
34
Proprietary Code Coexistence or Conflict?
Segregated Markets? Customized Software Mass
Market Software Competition Proprietary Code
Must Be Higher Quality Redefining
quality User friendliness, documentation,
support, backward compatibility. Installatio
n Costs Welfare Effects Constraints on
Proprietary Pricing Enforcing V C? Improved
Labor Markets
35
Limits to Open Source?
Does OSS Organize Some Tasks Better Than
Others? Building Core Code vs. Adding
Code vs. Fixes vs. Bug Reports Is OSS
limited by quality assurance? Is Modularity a
Free Parameter?
36
Limits to Open Source?
Many Projects are OSS in Name Only For every
successful OSS project, there are thousands of
unsuccessful ones. 80 of Projects have developers. 1 of Projects have 16
developers. 80 of Projects show no activity.
37
Beyond Open Source?
Complex Inventions GIS Drug Discovery