Models of Innovation: Linear Model

Models of Innovation

• Economists regarded technological innovation as a "black box," focusing solely on inputs and outputs
• Effective innovation policies rely on understanding how innovation occurs
• Two models exist: linear and chain-linked

The Linear Model

• This model goes through research, development, production and marketing
• Examples include medicines, JET, nuclear power, and mRNA

Linear Model: Science as the Source of Innovation

• Origins of linear model can be traced to the WWII science successes like the Manhattan Project
• Also, top-down (Fordist) managerial approaches of the era
• Key objective was supporting "big science" policies in the US and UK (1940s-60s)
• Linear model provides theoretical justification for public support for science
• Also, a key classification scheme for Research and Development (R&D) statistics

Drawbacks of the Linear Model

• This model lacks feedback and trials, so the innovation is not optimized, as shortcomings and failures are part of learning
• Learning by using feedback is important
• Boeing corrected the size of aircraft and the flight range of the 707 planes based on user feedback, making it successful
• Learning by doing is also neglected
• Exogenous view of technology means it is merely applied science
• Flow problems over wings opened a new scientific discipline: asymptotic perturbation theory
• Technology can create important inventions even when science is inadequate, like designing airplane wings without advanced asymptotic perturbation theory
• Central process of innovation is design instead of science
• Most innovation uses readily available knowledge; new research is only needed when existing knowledge is insufficient

State's Role in the Linear Model

• There are two market failures in knowledge production
• Suboptimal private provision of innovation
  • (Arrow, 1962)
• Suboptimal private provision of basic science
  • (Nelson, 1959)
    • Uncertainty in scientific knowledge application
• Countries take an active role in R&D, according to text

"Economic Welfare and the Allocation of Resources for Invention" (Arrow, 1962) - Assumptions

• Existing technology has constant average costs (c0)
• Non-drastic (incremental) process innovation is available
  • c1 < c0 AND pmonopoly(c1 ) > c0
• Innovation is introduced by spending a lump-sum R in R&D
• When introduced, the innovation is fully appropriable, having a patent of infinite length

"Economic Welfare and the Allocation of Resources for Invention" (Arrow, 1962) – Questions

• Who has the greatest incentive to innovate?:
  • Monopolist
  • Perfectly competitive firm
  • Public planner (state)

Arrow (1962) - Monopolist

• The Incentive is equal to the profit difference with and without innovation, assuming a discount rate of 1
• INCENTIVE = П(c₁) – П (c0)
• Profit equals total revenues minus costs
• A monopolist's innovation incentive is represented by area A on a graph

Arrow (1962) - Perfect Competition

• The price the innovative firm charges is equal to Co -ε
• Innovation incentive = P(c1) – 0 = A+B

Arrow (1962) - Public Planner

• Innovation incentive = Welfare (c₁) – Welfare (c0) =(W+Z) - Z area or A+B+C area

Arrow (1962) - Conclusions

• "A" represents the incentive for a monopolist to innovate, A + B is what motivates a firm in a perfect competition, and A + B + C for public planners
• Leaving investment in innovation to the private sector leads to underinvestment due to market failure
• Direct public investment in R&D and subsidies for private investment can remedy this

Arrow (1962) - Competitive Markets & R&D

• Firms in competitive markets have more incentives to invest in R&D

Nelson (1959): Knowledge and Uncertainty

• Three strong assumptions in Arrow (1962) are not realistic:
  • Innovation effort results in only one innovation
  • Complete appropriability of innovation
  • Absence of uncertainty
• Real-world R&D investments, especially basic research, do not align with these assumptions

Nelson (1959): Investing in Basic Research

• Basic research has uncertain results; large firms are better equipped to handle the risk
  • Big pharmaceutical firms developing drugs go through a "random screening" process
• One discovery can lead to several innovations
  • Gene editing allows for genetically modified plants, disease treatment, drug production (synthetic insulin), and genetically modified animals for research
• Basic research has very low appropriability:
  • Firms cannot patent scientific principles or formulas; only specific commercial applications
• Innovations take time:
  • The electromagnetism equations published in 1862 allow for today's use of x-rays
• Firms must meet a minimum R&D investment and budget to innovate

Nelson (1959): Basic Research Assumptions

• Uncertainty dominates R&D results
• Marginal benefits of basic research depend on innovation number, time frame, and sector
  • Larger, multi-product firms derive greater benefits
• The marginal cost of research is constant (equal to the wage of an additional scientist)

Nelson (1959): Large vs. Small Firms

• Large firms tend to benefit more from a given investment than small firms

Nelson (1959): Social Planner

• Social planners are incentivized to invest in basic research
  • Multi-product
  • No time constraints
  • Large budget for the development of the inventions deriving from basic research (HIGH MB)

Nelson (1959): Incentive for Basic Research

• Public planner > Monopolist (Large firm) > Competitive (small) firm
  • Public planners have largest incentives to invest in research
• Market failure occurs if investment in basic research is left to the private sector, resulting in underinvestment

Nelson (1959): Remedies to Market Failure

• Remedies to this issue:

1. State-funded basic research
2. Cooperative industry-public basic research
3. Prioritizing dynamic efficiency, allocating market power to large firms for basic research

What are the Remedies to Market Failure?: The State

• The state takes part in basic research through:
  • Big network of national institutions specialized in basic disciplines e.g Max-Planck-Institut (D), National Science Foundation -NSF (USA), Centre National de la Recherche Scientifique - CNRS (F), Centro Nazionale delle Ricerche -CNR (I)

State Remedies to Market Failure: Origins

• Origins can be traced back to the first half of the 20th century. Research of alternative and complementary facilities to universities
• Big national institutions dedicated to specific technologies e.g ENEA (ex atomic energy, now environment-energy…), Istituto Nazionale dei Tumori, Agenzia Spaziale Italiana
• Originated after WWII because of "Big Science" projects and linear innovation models
• In Anglo-Saxon and Scandinavian countries, Universities are strong. They are growing but were teaching institutions until the 19th century

Cooperative Basic Research Industry-Public

• The Bayh-Dole Act (1980) enabled universities and faculty to patent federally funded research
• The universities research has become appropriable and interesting for companies
• Creation of research units within universities designed to attract and accept commercial funds by promising the transfer intellectual property rights on resulting discoveries
• Universities took advantage of the biotechnology boom by signing massive research contracts, such as the $70 million deal between Hoechst and Harvard in 1984

Some Countries Prioritize Static Efficiency for Dynamic Efficiency

• Includes giving market power to big firms, so they can perform basic research
• AT&T's Bell Labs maintained a monopoly on markets versus production of knowledge as a public good until 1984

Trends Since 1950s: The Rise of Science and R&D Investment

• An explosion of funding was seen publicly between 1950 and 1960, and then a flat or decreasing trend (mini-boom in pre-crisis 2000s)
• Constant growth of research & development has been seen in businesses
• Emergence of East Asia in the process

Info from Official Statistics: Differences in intensity, composition of research

• Cross-country differences in R&D intensity and composition are very large due to:

1. Industrial structure (sectors, firms’ size …)
2. Science policies (high/low public financing, universities vs public labs…)
3. Military power ( defense industry)
4. Educational level of population

Chain-Linked Model

• An 'endogenous' view of science, along with S-T interaction
• Model is highly descriptive; hard to handle formally
• Origin:
  • Critique of US and UK science policies (1970s) based on the linear model
  • Critique of large firms’ approach to R&D
• Endogeneity of science and Complexity of S-T links
• Highlights how science connects to "hidden" innovation sources and challenges the sharp distinction between basic and applied research

Chain-Linked Model Diagram

• Process goes from:
• Scientific Research to Scientific Knowledge
• Scientific Knowledge to Radical Innovations
• Potential Market to Invention and/or Analytic Design to Distribute and Market
• The feed-back loop labeled f occurs during Redistribution and Design Phase, while feedback Loop labeled F occurs between Potential Market, Analytic Design and and produce

Chain-Linked Model: Central-Chain Innovation

• In terms of the "First Part", The main path is referred to as central- chain innovation, is indicated with the letter 'C'

Chain-Linked Model: Economic vs. Technological Side

• The economic side is as important as the technological side
  • Concorde -> brilliant technological achievement but also a costly commercial failure -> half of the time but the fuel per passenger costs 15 time more than a 747
  • Solar energy -> will be appealing when it will be closer to the cost of other sources"

Chain-Linked Model: Analytic Design and Invention

• Invention: it is a new mean for achieving some functions not obvious to someone skilled in the prior art (e.g., gene editing technique, transistor, …)
• Analytic design: innovation by re-configuration-> the technology of the base components is already known but put together in a new configuration

Chain-Linked Model - Second Part

• The link from science to innovation is not only in the beginning of the innovation process, but rather extend alongside all the development process

Chain-Linked Model - Arrows

• Use of science occurs in two stages:
  • In the face of a technical problem, first use known science from stored scientific knowledge (asking to an expert) If stored knowledge fails, turn to scientific research via university collaboration
• Research is more costly and time consuming

Chain-Linked Model - Radical & Product Innovation

• Arrow D signifies radical innovations, such as lasers, semiconductors, and genetic engineering
• Arrow I reflects the feedbacks to science from product innovation Microscope -> modern medicine
  • Telescope -> astronomy