View
215
Download
0
Embed Size (px)
Citation preview
L. Orsenigo, OU, October 24th, 2007
A HISTORY FRIENDLY MODEL
OF THE CHANGING VERTICAL SCOPE OF FIRMS IN
OF THE COMPUTER AND SEMICONDUCTOR INDUSTRIES
F. Malerba, R. Nelson, L. Orsenigo and S. Winter
L. Orsenigo, OU, October 24th, 2007
Understanding the determinants of specialization and vertical integration in related industries
in uncertain and dynamic environments,
characterized by technological discontinuities.
Major factors: capabilities, technical change and market size
Co-evolutionary processes
L. Orsenigo, OU, October 24th, 2007
1. The history of the industries
2. The conceptual framework
3. History-friendly models
4. The main structure of the model
5. The baseline simulation
6. Counterfactual analyses
7. What did we learn?
L. Orsenigo, OU, October 24th, 2007
Focus on:
- the history of the American industry
- standard components
- vertical integration of system producers into component R&D and production (upstream vertical integration)
We will not discuss
-Second sourcing
-Intermediate organizational forms (such as R& networks and partial integration)
L. Orsenigo, OU, October 24th, 2007
A BRIEF HISTORY OF THE COMPUTER AND SEMICONDUCTOR INDUSTRIES (1950-1985)
Three main periods examinedThree main periods examined::
a. Mainframes and transistors as their main SC components
Emergence of IBM as monopolist in mainframes
b. Introduction of the integrated circuit (IC)
c. Introduction of microprocessors (MP) and birth of the personal computer (PC) industry
L. Orsenigo, OU, October 24th, 2007
Vertical integration and specialization patterns
Early in the history of the computer industry, most computer producers were not vertically integrated.Then some became vertically integrated.
With the introduction of the integrated circuit, IBM was fully integrated into IC, because of coordination advantages (IC embedded system elements), fears of leakages of strategic information and security of supply reasons.
With the introduction of microprocessors (MP), IBM and other vertically integrated producers dis-integrated from the large scale production of standard semiconductors and moved to specialization, because they faced a major technological discontinuity and quite large and competent MP firms.
L. Orsenigo, OU, October 24th, 2007
From history:
-COMPETENCES:
accumulation of competences in specific technological or market domains (IBM, Intel)
coordination and integration capabilities;
-TECHNOLOGICAL DISCONTINUITIES
if major, there is destruction of established competences
-SIZE OF FIRMS (which affects R&D expenditures and security of supply)
-SIZE OF DEMAND
-MARKET STRUCTURE of two vertically related industries
L. Orsenigo, OU, October 24th, 2007
A capability-based, dynamic A capability-based, dynamic appreciative theory appreciative theory
of vertical integration and specialization of vertical integration and specialization
1. Competence accumulation in specific technological and market domains
2. Competence destroying technological change
3. Coordination and integration capabilities
Ref: Malerba-Nelson-Orsenigo-Winter ICC 1999
Jacobides-Winter SMJ 2005
L. Orsenigo, OU, October 24th, 2007
Capabilities take time to develop
Decisions to specialize and vertically integrate are not symmetrical
The distribution of capabilities among all industry participants are relevant
Market selection amplifies the effects of capabilities on the vertical scope of firms
The identity of firms affects the development of capabilities
L. Orsenigo, OU, October 24th, 2007
Let’s consider a system industry
Capabilities and technical change interact in specific ways.
When technical change proceeds along a well defined trajectory, vertical integration of system producers will take place if component capabilities and system capabilities are related in terms of knowledge or if the capabilities of coordinating and integrating components and systems into new products are important.
When technological change in systems has a major discontinuity, vertical integration may take place because of the misalignment of the capabilities of system producers and component producers(for example, because of the a slow responsiveness in the development of new capabilities in new components by existing component producers) .
L. Orsenigo, OU, October 24th, 2007
The opposite holds for specialization
When technical change proceeds along a well defined trajectory, specialization of system firms may take place when component and system companies are quite different in their knowledge and competences.
When technological change in components has a major discontinuity, specialization of system firms may take place if totally new capabilities are needed for developing new components and entry is relevant.
L. Orsenigo, OU, October 24th, 2007
However also MARKET SIZE may become a major factor affecting vertical integration and specialization.
Think of the market for PC and the derived demand for semiconductor
Or think of American military procurement
For example, in addition to a market for components related to large systems, an additional large external markets for components may support the growth and the specialization of component suppliers, which can therefore develop their capabilities and become quite innovative compared to vertically integrated system firms.
Relatedly, the absence of external markets, make components producers quite dependent on the decision to integrate or specialize by the existing system industry.
L. Orsenigo, OU, October 24th, 2007
Also the emergence of a powerful MONOPOLIST may have a role in integration and specialization
High CONCENTRATION in the system industry may lead to vertical integration in periods of incremental technical change if system producers are confronted with an ATOMISTIC component industry (because of security of supply reasons). On the contrary, a FRAGMENTED system industry confronted with a MONOPOLISTIC component industry may have a drive for specialization (because of the large size of the R-D expenditure of the component monopolist).
L. Orsenigo, OU, October 24th, 2007
All these factors are closely related
and cannot be examined in isolation.
So vertical integration and specialization are the result of the CO-EVOLUTION of competences, technological change, and the size of markets.
L. Orsenigo, OU, October 24th, 2007
HISTORY FRIENDLY MODELS (HFM)
HFM intend to enhance the understanding of particularly interesting and important phenomena
HFM are simulation models
HFM are models which aim to capture in stylised form qualitative theories about mechanisms and factors affecting industry evolution and technological advance
They do not aim to match the quantitative values observed
They do not specify the model parameters as close as possible to actual empirical values
L. Orsenigo, OU, October 24th, 2007
HFM aim to explore whether particular mechanisms and forces built into the model can generate (explain) the patterns examined
HFM are guided by verbal explanations and appreciative theorizing
HFM are “evolutionary”: boundedly rational agents; behavior guided by routines; learning and capabilities as key variables; historical processes; competition and selection
L. Orsenigo, OU, October 24th, 2007
HFM establish both runs that match the qualitative features of historical patterns and runs that do not match the historical patterns (counterfactuals analyses).
They allow to explore the role of specific factors and mechanisms in industry evolution
They tackle several key questions in industrial organizations and economics of innovation in a dynamic and empirically-grounded setting
L. Orsenigo, OU, October 24th, 2007
THE MODEL
L. Orsenigo, OU, October 24th, 2007
Computers have a mix of chacteristics: cheapness and performance
Computers use semiconductor (SC) components
SC components are sold to computer firms and to an external market
At the beginning of the evolution of the computer industry: SC component technology- transistors- makes possible to have mainframes which are sold to big users (which are more interested in performance than in cheapness)
First technological discontinuity in SC components: integrated circuits (IC) Entry of new SC firms producing IC
Mainframe producers adopt IC in mainframes
Second technological discontinuity in SC components: microprocessor (MP)
Entry of new SC firms producing MP
MP are used in mainframes and in the same time they open a new market: personal computers (PC). The new PC market appeals to a new set of consumers -individuals- who are more interested in cheapness than in performance.
L. Orsenigo, OU, October 24th, 2007
SPECIALIZATION AND INTEGRATION DECISIONS
In the model
VERTICAL INTEGRATION decision is led by:
- the relative size of the computer firm compared to the largest SC component producer
-the age of the SC component technology
SPECIALIZATION decision is led by:
-Comparison between the quality of SC components produced in-house and the quality of SC components available on the market
L. Orsenigo, OU, October 24th, 2007
COMPUTERS
Merit of design M of computers
1
,,, 1 Sti
Ctiti MMAM
C= components
S= systems
L. Orsenigo, OU, October 24th, 2007
DEMAND FOR COMPUTERS
11,
1, )1(
titi sML
iti
titi L
L
,
,,Pr
1,,, tititi zwM
w = cheapness z = performance s= market share
L. Orsenigo, OU, October 24th, 2007
DEMAND FOR COMPONENTS
21,
2, )1(
ticC
ti sML
Cti
CtiC
ti L
L
,
,,Pr
L. Orsenigo, OU, October 24th, 2007
FIRMS’ BEHAVIOUR AND TECHNICAL PROGRESS
PROFITS:
PRICE:
R&D OF INTEGRATED FIRMS:
FIRM DRAWS:
MEAN OF NORMAL DISTRIB:
PUBLIC KNOWLEDGE:
v
Rd ti
ti,
,
tiCti
Cti RfcmR ,,,
Kttiti KhMh )1(1,,
tititititi oMpM ,,,,,
)1(,, mop titi
k
tK
Kt tctn
eK k1
1lim
O = cost
L. Orsenigo, OU, October 24th, 2007
VERTICAL INTEGRATION
ti
titi V
VbIntegrateob
,
,, 1
)(Pr
2
,
1
, 1;min
Ct
titi q
q
g
AV
A =time from each discontinuity q = size
L. Orsenigo, OU, October 24th, 2007
SPECIALIZATION
0,
maxmax
,
,, C
ti
Cti
Ct
ti M
MMZ
ti
titi Z
ZASpecializeob
,
,, 1
)(Pr
L. Orsenigo, OU, October 24th, 2007
EXIT
titi seleE ,, )1(
Exit if E i,t < E
L = inverse of the number of firms active in the market at the beginning of the simulation
E = constant
L. Orsenigo, OU, October 24th, 2007
THE HISTORY-FRIENDLY SIMULATION
- Mainframes and the emergence of IBM
IBM vertically integrates in SC components
- Integrated circuits (IC):
Entry of new SC firms
IBM remains vertically integrated
- Microprocessors MP:
Entry of new MP firms and a large external market for MP
- A new computer market (PC) emerges for different users (individuals)
PC producers remain specialized and buy MP
- A dominant microprocessor firm emerges in SC industry
IBM dis-integrate
L. Orsenigo, OU, October 24th, 2007
History Friendly Simulation
Figure 1a: Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 17 33 49 65 81 97 113 129 145 161 177 193 209 225 241
MF
PC
Cmp
Figure 1b: integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 19 37 55 73 91 109 127 145 163 181 199 217 235
mf
pc
L. Orsenigo, OU, October 24th, 2007
TESTING THE MODEL:COUNTERFACTUALS
1. Does lack of external markets lead to more vertical integration?
2. Do no demand lock-ins in mainframes lead to more specialization ?
3. Do no demand lock-ins in semiconductors lead to more vertical integration ?
4. Does a minor technological discontinuity in microprocessors lead to more vertical integration?
L. Orsenigo, OU, October 24th, 2007
No external market for SC
Figure 2a: Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 20 39 58 77 96 115 134 153 172 191 210 229 248
MF
PC
Cmp
Figure 2b: integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 19 37 55 73 91 109 127 145 163 181 199 217 235
mf
pc
L. Orsenigo, OU, October 24th, 2007
No lock-ins effects on Mainframes
Figure3a: Herfindahl index
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
1 15 29 43 57 71 85 99 113 127 141 155 169 183 197 211 225 239
MF
PC
Cmp
Figure3b: integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 20 39 58 77 96 115 134 153 172 191 210 229 248
mf
pc
L. Orsenigo, OU, October 24th, 2007
No lock-ins effects on Microprocessors
Figure 4b: integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 20 39 58 77 96 115 134 153 172 191 210 229 248
mf
pc
Figure4a: Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 15 29 43 57 71 85 99 113 127 141 155 169 183 197 211 225 239
MF
PC
Cmp
L. Orsenigo, OU, October 24th, 2007
MP is not a major discontinuity
Figure 5a: Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 20 39 58 77 96 115 134 153 172 191 210 229 248
MF
PC
Cmp
Figure 5b: integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 18 35 52 69 86 103 120 137 154 171 188 205 222 239
mf
pc
L. Orsenigo, OU, October 24th, 2007
Antitrust
• Antitrust intervenes when a firm has a market share > 90% and breaks the monopolist in two companies.
• No effect: given strong lock.in effects, a new monopolist emerges very quickly
• Antitrust intervenes also by reducing lock-in effects• Effects:
– competition in mainframes – Slower vertical integration– Concentration grows in components market in the age of
transistors and integrated circuits– Competition and specialization in all markets in the
microprocessor era– No effects on the rate of technical change
L. Orsenigo, OU, October 24th, 2007
Antitrust: simulation
Herfindahl Index
00,10,20,30,40,50,60,70,80,9
1
1 19 37 55 73 91 109 127 145 163 181 199 217 235
MF
PC
Cmp
integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 18 35 52 69 86 103 120 137 154 171 188 205 222 239
mf
pc
L. Orsenigo, OU, October 24th, 2007
Selective public procurement
• In the age of transistors and integrated circuits the best component firm gets additional sales from the external market
• Concentration increases in component market, but decreases sharply in the era of microprocessors
• Slower integration
• No effects on the rate of technical change
L. Orsenigo, OU, October 24th, 2007
Selective public procurement: simulation
Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 19 37 55 73 91 109 127 145 163 181 199 217 235
MF
PC
Cmp
Integration ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 18 35 52 69 86 103 120 137 154 171 188 205 222 239
mf
pc
L. Orsenigo, OU, October 24th, 2007
Investment in basic research
• Increase in the rate of growth of public knowledge
• Faster integration in transistors and integrated circuits; increase in integration in the microprocessors era: better computers induce higher demand and faster firms’ growth
• Enormous increase in the rate of technical change
L. Orsenigo, OU, October 24th, 2007
Investment in basic research: simulation
Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 19 37 55 73 91 109 127 145 163 181 199 217 235
MF
PC
Cmp
int ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 18 35 52 69 86 103 120 137 154 171 188 205 222 239
mf
pc
L. Orsenigo, OU, October 24th, 2007
Favouring circulation of knowledge
• Technical change is less cumulative at the firm level
• No effect on concentration in mainframes (high bandwagon)
• No change in integration
• Slight decrease in the rate of technical change in mainframes
L. Orsenigo, OU, October 24th, 2007
Favouring circulation of knowledge: Simulation
Herfindahl index
0
0,2
0,4
0,6
0,8
1
1,2
1 19 37 55 73 91 109 127 145 163 181 199 217 235
MF
PC
Cmp
int ratio
0
0,2
0,4
0,6
0,8
1
1,2
1 18 35 52 69 86 103 120 137 154 171 188 205 222 239
mf
pc
L. Orsenigo, OU, October 24th, 2007
The creation of open standards in computers leads to the
emergence of concentration in components • open standards and elimination of the bandwagon
effects reduces concentration in all industries. • lower concentration in mainframes entails a reduction in
vertical integration and an increase in the demand for microprocessors from the previously integrated large mainframe producers.
• a de-facto standard concerning the interfaces between components arises, showing itself in the form of the emergence of a bandwagon in component demand.
• This market-driven increase in bandwagon in microprocessors created by the additional demand from large specialized computer producers generates concentration in the microprocessor markets.
• unintended consequence of policy.
L. Orsenigo, OU, October 24th, 2007
The creation of open standards in computers leads to the emergence of concentration in components
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 248
MF
PC
CMP
Figure 1 – Unintended consequences: open standards and concentration. Herfindahl index
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 248
MF
PC
L. Orsenigo, OU, October 24th, 2007
Antitrust policy in computers leads to the emergence of a monopolist in a related system market and the disappearance of
a the merchant component industry
• an antitrust policy breaks the mainframe monopolist in two• one of the two producers diversifies into personal computers, • a new large producer enters this industry with a relevant brand
name. • Because of its size, reputation and marketing capabilities, this
producer is able to increase the level of the bandwagon effect in the personal computer industry and become the leader in this industry. This is a first unintended consequence of public policy.
• In addition, the new personal computer monopolist may become vertically integrated into microprocessors.
• This may lead to the disappearance of the microprocessor industry, if there are no other external markets for semiconductors. This is the second unintended consequence of the policy
L. Orsenigo, OU, October 24th, 2007
Antitrust policy in computers leads to the emergence of a monopolist in a related system market and the disappearance of
a the merchant component industry
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 248
MF
PC
CMP
Figure 1 – Unintended consequences: antitrust, diversification and disappearance of the component industry:
Herfindahl index
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 248
MF
PC
Figure 2 – Unintended consequences: antitrust, diversification and disappearance of the component industry.
Integration ratio
L. Orsenigo, OU, October 24th, 2007
CONCLUSIONS
1. The history-friendly model is able to reproduce the main stylized facts of competition and vertical integration in the computer and semiconductor industries
2. Specialization and vertical integration are functions of competences, technological change and size of markets
3. Co-evolutionary processes are relevant
4. Specific conditions affect the vertical scope of system firms:
- Size of external markets
- Magnitude of technological discontinuities
- Lock-ins effects in demand
5. Policies: inefficacy propositions and unintended consequences