Transcript Slide 1
Economics 124/PP 190-5/290-5 Innovation and Technical Change Science, invention, and innovation Prof. Bronwyn H. Hall UC Berkeley Today Science, invention and innovation The linear model Deviations from the linear model Technology driving science Learning by using Chance and unexpected innovations Lessons for policy Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 2 The linear model of innovation A useful conceptualization, but not the whole story. The idea: science base → basic research → applied research → invention → prototype → development → commercialization → diffusion → technical progress → economic growth Sometimes the entire process in red is referred to as innovation Which stages need funding, and how? Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 3 Example – new drug Basic research – microbiology, etc. Applied research – screening compounds in test tubes; testing on animals Invention – successful in laboratory Development – Phase I and II clinical trials Commercialization – packaging; marketing; dosage info Diffusion – spread throughout the patient/doctor population Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 4 Example – new software Basic research – mathematics, queuing theory Applied research – cryptography, sorting algorithms, data storage systems Invention – idea of program, design, basic features Development – programming, detailed specifications, alpha testing Commercialization – beta testing, marketing, sale Diffusion – adoption by consumers; large market share Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 5 Modifying the linear model Importance of backward links (reverse the arrows) Commercialization and diffusion → new innovation & development Invention/innovation → science base/basic research Rosenberg emphasizes this point in a series of papers “How exogenous is science?” (1981) “Learning by using” (1978) Both published in Inside the Black Box (CUP 1982) Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 6 Backward links in the linear model "How Exogenous is Science?" how applied research and innovation have yielded new scientific knowledge and created new scientific fields, both accidentally and purposefully => feedback from applied research, innovation, and development to the science base. "Learning by Using" a term modeled on learning by doing describes how products are improved and developed in both embodied and disembodied ways, based on experience of the product in use => feedback from diffusion to development Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 7 How exogenous is science? exogenous means determined outside the system in this case, the innovation system Technological knowledge often precedes scientific knowledge Scientific progress can be an accidental byproduct of searching for an answer to a technological puzzle A technological discovery can dictate the direction in which subsequent scientific research should go Improvements in instrumentation (technology) A major determinant of subsequent scientific progress nanotechnology (1910) and the electron microscope Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 8 Science from technology (1) Inventor/s cientist Technological source Years Scientific field developed Toricelli improved pump – explored the weight of the atmosphere 1600s Atmospheric/pressure science/barometer Watt/Carnot steam engines 1830s Thermodynamics Pasteur wine industry/ fermentation 1850s Bacteriology/germ theories Perkin /Hoffman Synthesis of mauve, first aniline dye 1870s Organic chemistry Wilm Bessemer process; age-hardening of duraluminum 1850- Metallurgy/ 1900s materials science Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 9 Science from technology (2) Inventor/ scientist Technological source Years Scientific field developed Davisson vacuum tubes – patterns of emission from nickel crystal due to electrons 1920s Wave nature of matter/ Nobel prize 1937 Jansky/ Bell labs radio noise 1932 Radio astronomy (star noise) Townes/ Bell labs Laser technology for optic fiber cables 1950s Optics resurgence Shockley Transistor/ semiconductor 1948 Solid state physics Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 10 Innovation and learning During the R&D process Knowledge concerning laws of nature (basic R) Knowledge with useful applications (applied R) Knowledge directed towards optimal design characteristics and satisfying consumer wants (development) After the R&D process During manufacturing – learning by doing During the use of the product – learning by using Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 11 Learning by doing Widespread phenomenon in the repeated manufacture of a good (airframes, chemicals, semiconductors) First measured for aircraft in the 1930s-1940s labor = N-1/3 where N=N airplanes produced Observed in a number of industries by the Boston Consulting Group – plotted learning curves with downward slope ~ 0.3 A major feature of semiconductor manufacturing as the number of rejected chips falls over time Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 12 Learning in chemicals 25000 Unit Cost 20000 15000 10000 5000 .15 .2 .25 .3 .35 .4 .45 .5 .55 Cumulative Production Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 13 Learning by using Technological change does not end after the technology is diffused Technologies continue to improve due to feedback from use and users software; skateboards Some improvements embodied learning how to stretch a Boeing 747 Some disembodied learning the maintenance frequency necessary for aircraft learning that a drug good for one condition actually works for another Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 14 Learning by using Performance of complex capital goods not fully understood until they are used Technological knowledge required highly specialized, includes user knowledge Product differentiation valuable to users, can be achieved by them e.g., skateboard innovations from users Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 15 Aircraft example Embodied learning Uncertainty in aircraft design, plus caution in first use As time goes by, experience leads to stretched aircraft, larger payloads Disembodied learning Extensive maintenance and overhaul requirements of jet engines Over time, service intervals lengthened Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 16 Boeing 737-100,737-600/900 Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 17 Boeing 747-100,747-300,747-500 Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 18 Uncertainty and chance History of innovation replete with examples of discoveries that Were a side effect of a completely different investigation, as in technology>science examples were unappreciated at the time they were made (consequences or usefulness unforeseen) Examples: laser, radio, computer Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 19 Examples of forecasting failure Laser (Light Amplification by Stimulated Emission of Radiation) invented by Townes at Bell Labs around 1960 now used in navigation, precision measurement, chemical research, surgery, compact discs and printing most important and widespread use is probably fiber-optic cable for telecommunications But…..lawyers at Bell labs did not apply for a patent, thinking it not relevant for their industry, which was the telephone industry! Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 20 Examples of forecasting failure Radio invented by Marconi to be useful when wire communication impossible, such as ship-to-ship, that is, narrowcasting, not broadcasting Computers Watson, Sr. (head of IBM) saw a need for only one computer to solve all the world’s scientific problems In 1949, IBM forecast world demand at 10-15 computers (ENIAC contained 18,000 vacuum tubes and was 100 feet long, so this is understandable) Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 21 Some lessons for policy Macro-inventions (scarce ideas?; radical innovation) Unexpected sources and consequences May require a broad science base Micro-inventions (well-known needs?; incremental innovation) More predictable Often a result of natural evolution of a technology Easier to pay for Distinction does not necessarily correspond to the increment in economic welfare (e.g. malaria vaccine) Fall 2004 (C) B H Hall Econ 124/PP 190-5/290-5 22