Institutional foundations of scientific progress: Implications for collaboration & participation Fiona Murray Associate Professor Technological Innovation & Entrepreneurship Group MIT Sloan School of Management Presentation to COMMUNIA 2009

Why is there an emphasis on collaboration in science today? It is not “desirable” in itself therefore it must be regarded as a means to desirable ends.

Typical “working assumptions” regarding collaboration in science Collaboration allows for more rapid progress Collaboration allows for broader participation – democratization of science Collaboration allows for the production of more diverse research

Alternatively, re-designing scientific institutions may be more effective than re-engineering collaboration! Current problems in science emerge from pressures on informal institutions that traditionally supported rapid, diverse scientific progress. Re-designing the institutional arrangements for science would be a more effective and longer lasting approach.

Four Characteristics of Scientific Knowledge Its production is a step-by-step process Outputs from one step are inputs into the next (Aghion & Howitt 2000) It is multi-faceted Key outputs include information, materials, know-how, methods (Collins 1974) It is non-rival in use One idea can be an input into many follow-on experiments (Arrow 1952, Rosenberg 1986) Its follow-on value can be uncertain Wide ranging perspectives on what the most interesting possible direction (Shane 2001) Image is of the first laser – developed by Bell labs. Lawyers were reluctant to patent because they did not think it had any application to telephony and could not think of any other uses! Lasers are used for navigation, precision measurement, and chemical research. In surgery, detached retinas, a frequent cause of blindness, now are repaired with lasers on an outpatient basis. Lasers are gynecologists' instrument of choice in many surgical procedures. In the textile industry, they are used to cut material, speedily and accurately. Lasers are employed in millions of households for the high-quality reproduction of music recorded on compact discs. The laser's most profound impact so far has been in telecommunications, where, with fiber optics, it is revolutionizing transmission. In 1966, the best transatlantic telephone cable could carry 138 conversations simultaneously between Europe and North America. The first fiber optic cable, installed in 1988, could carry 40,000; today's cables, nearly 1,500,000. USA Today (Society for the Advancement of Education),  May, 1995  by Nathan Rosenberg

Implications for scientific research For follow-on research to be rapid, productive, inclusive & cumulative we must ensure that scientific knowledge outputs are widely available: => Access to multiple dimensions of a given knowledge output Access to many follow-on innovators not merely a small group Access terms unrestrictive – should not prescribe the path for follow-on innovation i.e. must allow follow-on innovators to select their chosen line (Aghion, Dewatripont & Stein 2008)

Meeting these conditions requires a complex of scientific institutions Mere production of knowledge does not guarantee its use by the next generation (Rosenberg 1984; David & Dasgupta 1994; Mokyr 2002): Complex of institutions is needed to support disclosure and exchange of knowledge outputs for follow on researchers. Disclosure –what is disclosed and thru what means? Access – under what terms is knowledge provided for replication, validation & accumulation? Rewards – what are the rewards for disclosure & access

Informal peer-to-peer institutional complex & few formal institutions Traditional scientific progress was supported by informal institutional complex Small elite communities Grounded in close personal relationships e.g. Erdos Old World Informal peer-to-peer institutional complex & few formal institutions Knowledge exchanged and accumulated peer-to-peer under shared norms & culture Informal mechanisms to deal with competition, lack of reciprocity etc.

Letter were another critical form of exchange that operated within the informal structures Multiple letters exchanged between Watson & Crick, Max Delbruck, Linus Pauling in the race to the structure of DNA Letter from Watson & Crick to Delbruck, March 12 1953

Scientific materials in private collections were accessible to those participating in informal institutions FLIES gathered, bred & shared under the watchful eye of Thomas Hunt Morgan (“Lord of the Flies” (Kohler 1988) & his former post-doc Sturtevant KEY LIMITATION Local or at least limited to a few people Extremely costly High barriers to entry Limitation s become manifest in a few critical scientific scandels A.H. Sturtevant in the Drosophila stock room. Courtesy of the Caltech Archives ©

Clubs & societies played a central role in the traditional informal institutions “The origins of the Royal Society lie in an "invisible college" of natural philosophers who began meeting in the mid-1640s to discuss the ideas of Bacon”. While democratic on some levels, Shapin (1984) vividly described participation was closely guarded & tied to being a “gentleman”

Informal institutional complex functioned effectively to allow for speed, participation & diversity of follow-on research in a small scientific community SUPPLY-SIDE STRAINS DEMAND-SIDE STRAINS Knowledge outputs growing in complexity Knowledge outputs increasingly costly to maintain; hard to standardize & validate New rules e.g. Bayh-Dole requires suppliers to triage the terms of access Increase in size of the scientific community Growing participation in science from nations not traditionally part of invisible college Growing diversity of participants – e.g. industry demand for scientific knowledge

Re-engineering collaboration is a widespread response to this crisis It forces researchers with key knowledge outputs to work with a more diverse groups of researchers It requires them to work on different types of follow-on projects By collaborating it may increase speed (although the evidence is weak) BUT – it reduces the freedom of researchers which decreases a powerful motivating incentive (Aghion et al. 2008) INSTEAD Why not focus on the root cause of the problem – the limits of informal institutions

Alternative Response: Re-design institutional complex to establish a complex of more formal institutions ICSD: All crystal structures contained in the database have been carefully evaluated and checked for quality by our expert editorial team.  We continuously extract and abstract the original data from over 80 leading scientific journals and additional 1,900 scientific journals. The most important source of new data for GenBank® is direct submissions from scientists. GenBank depends on its contributors to help keep the database as comprehensive, current, and accurate as possible. Over 16,000 researchers in more than 60 countries rely upon JAX® Mice. They are the most published and well characterized models in the world

Key Question Do these formal institutions actually increase efficiency of exchange, broaden participation & increase diverse of follow-on research In other words… if we design these institutions correctly, can we accomplish many of the goals that we strive for in our funding of science – and do so in such as way as to reduce the need to re-engineer collaboration & thus increase scientific freedom?

Empirical Analysis of these formal institutional changes is challenging Need to define the counterfactual: What type of innovation would we expect to take place under different institutional conditions? Need systematic measures of follow-on: Can we measure the follow-on innovation we care about? Need “experiments”: What sorts of settings allow us to examine institutional arrangements that vary independent of variations in the underlying type of knowledge? i.e. identification is key => we need to distinguish between selection & marginal effects

Analyzing Effectiveness of Formal Scientific Institutions Pre-period institutional setting Post-period institutional setting FCjt Exogenous SHIFT in institutional arrangements Measure citations before & after to estimate impact of SHIFT Publication FCjt FCjt Treatment group Publication FCjt FCjt Publication Meaning of citations: A citation in Patent X to Patent Y mean that “X represents a piece of previously existing knowledge on which Y builds”. In practice a citation has strong competitive implications – it restricts the claims of patent Y to only those things that extend beyond the scope of Patent X. Paper Citations: A Paper citation is more ambiguous in its meaning. The citation of a paper in a patent has been taken to mean that this public knowledge is being built upon by the inventor. Who again claims only the ideas above and beyond the innovation. It has been suggested that this references relate to the basicness of the patent and its so-called “science-linkage”. While this may be true it is not necessary to this analysis. References to science in other papers has been widely studied not only in bibliometric analysis of knowledge creation but also by sociologists of science who have suggested that the process is quite political, favors friendship and so on. Thus the two types of citation are functionally but not structurally equivalent. Nonetheless in simple terms I take them as a narrow definition of the community first of commerce and second of science. FCjt Control group Publication Plot forward citations over time as a measure of scientific knowledge accumulation building on a “piece of knowledge”

Analyzing Effectiveness of Formal Scientific Institutions Pre-period institutional setting Post-period institutional setting FCjt Use forward citations to capture a variety of features of follow-on knowledge production: Level (pre- and post-) Diversity of participation (new versus already contributing authors or universities) Diversity of ideas (new versus already used key words or journals) Publication FCjt FCjt Treatment group Publication FCjt FCjt Publication Meaning of citations: A citation in Patent X to Patent Y mean that “X represents a piece of previously existing knowledge on which Y builds”. In practice a citation has strong competitive implications – it restricts the claims of patent Y to only those things that extend beyond the scope of Patent X. Paper Citations: A Paper citation is more ambiguous in its meaning. The citation of a paper in a patent has been taken to mean that this public knowledge is being built upon by the inventor. Who again claims only the ideas above and beyond the innovation. It has been suggested that this references relate to the basicness of the patent and its so-called “science-linkage”. While this may be true it is not necessary to this analysis. References to science in other papers has been widely studied not only in bibliometric analysis of knowledge creation but also by sociologists of science who have suggested that the process is quite political, favors friendship and so on. Thus the two types of citation are functionally but not structurally equivalent. Nonetheless in simple terms I take them as a narrow definition of the community first of commerce and second of science. FCjt Control group Publication

Specific example of formal institutional arrangements A tale of three (blind, obese, diabetic, epileptic…) mice engineering technologies…. …setting to explore impact of changes (negotiated by NIH) that allowed for both greater formal access (via JAX) and lower IP restrictions Onco transgenic mouse technology Cre-lox mouse technology Knock-out mouse technology

EMPIRICAL APPROACH “Mouse-articles” - Sample of mice linked with specific scientific articles –Cre-lox, Onco, Knock-Out& Spontaneous (1983-98) Mouse Genome Informatics (MGI) database catalogs over 13,000 mice & links each mouse to an original publication in a scientific journal – mouse-articles Cre-Lox (52), Oncomouse (160), Knock-Out (2171), Spontaneous (255) Citations to mouse-articles in other scientific publications observed over time PubMed for information about mouse-articles & ISI Web of Science SCI for citations “Shifts” to the openness of Cre (1998) & Onco (1999) mice–after their publication - but not to Knock-out or Spontaneous => observe the citation of mouse-articles in both the pre- and post- shift period and compare the use of “treatment” and “control” group => evaluate the impact of openness on the rate & nature of follow-on research

Analysis: Effectiveness of Formal Institutions for Changing Access to Research Mice Neg. Binomial Last Authors Key Words Annual Citations with New Last Author Annual Citations with Old Last Author Annual Citations with New keywords Old keywords Post Shock 1.380*** 1.14 1.260*** 0.977 Conditional Fixed Effects for Article, Margin-Age and Margin-Calendar Year, Window Effects 26% Boost After NIH Agreement formalizes Access & lowers IP The impact of institutional change concentrated in citations by “new” last authors and in papers using new key words Robust to “New Institution” v.“Old Institution”, Reprint Authors, Journals etc. Murray, Aghion et al., 2009 Murray, Aghion et al., 2009

Analysis: Results for Biological Resource Center Deposit Neg. Binomial Model Dep. Var. Annual Forward Citations BRC-Article,Post-Deposit 2.35 Article FE Y Age FE Calendar Year FE 135% Boost After Deposit Evidence for the positive impact of material deposit and lower cost open access on scientific accumulation Furman & Stern 2008

US-only collaboration Analysis Impact of stem cell policy changes (restricting formal institutions supporting access to cells & funding) Author Type No collaboration US-only collaboration US & EU collaboration Top 50 1.553 (0.431) 1.084 (0.203) 3.180 (1.781)** US-Not Top 50 0.986 (0.228) 1.421 (0.235)** 5.137 (2.512)*** Article FE Y Age & Year FE 414% Increase in International collaboration- based citations after policy Evidence for shift in entire collaborative structure in human embryonic stem cell research after government policy changes restrict access to cell lines through formal institutional arrangements & formal funding channels. Furman, Murray & Stern 2008

Conclusions Collaboration is increasingly imposed to solve a set of problems with the nature of scientific research - The speed - The diversity of participants - The breadth of follow-on paths pursued These challenges have emerged because the informal institutional complex that traditionally supported the speed, diversity & breadth of scientific research are strained Alternative to re-engineering collaboration is re-designing institutions – emphasizing formal institutions allowing for rapid & simple access to key research outputs & thus - Speeding-up follow-on innovation - Increasing diversity of participants by decreasing reliance on informal ties - Increasing the breadth of follow-on paths pursued by limiting restrictions

Open Questions Global Access to Scientific Knowledge Can the shift from informal to formal institutional complex really bring global equity in participation? Is collaboration still required at the emerging “margins” to jump start participation? What are the strategies to get to the knowledge frontier with and without formal institutions?

Questions or comments? fmurray@mit.edu Thank you! Questions or comments? fmurray@mit.edu

SELECTED REFERENCES P. Aghion, J. Kolev, F. Murray & S. Stern (2009). Of Mice & Academics: Examining the Effect of Openness on Innovation. NBER Working Paper. F. Murray (2008). “The Oncomouse that Roared: Resistance & Accommodation to Patenting in Academic Science”. Forthcoming in Amercian Journal of Sociology F. Murray and S. O’Mahony (2007). “Re-conceptualizing the Institutional Foundations of Cumulative Innovation.” Organization Science.  F. Murray and S. Stern (2007). “Do formal Intellectual Property Rights Hinder the Free Flow of Scientific Knowledge? Journal of Economic Behavior and Organization.  F. Murray (2007). “The Stem Cell Market: Patents & the Pursuit of Scientific Progress.” New England Journal of Medicine, Vol. 356:23, pp. 2341-2343.  K. Jensen, C. Jinan, F. Murray (2007). A Simple Method to Improve Life Science Patent Searches using the Cyber infrastructure of the NIH. First Tuesday. K. Huang and F. Murray (2007). Golden Helix or Tangled Web: The Impact of Gene Patents on the Diffusion of Genetic Knowledge. Working Paper. F. Murray and D. Spar (2006). “China and Stem Cell Research: Bit-Player or Powerhouse?” New England Journal of Medicine, Vol. 355:12, pp. 1191-1194.  K. Jensen and F. Murray (2005). “The Intellectual Property Landscape of the Human Genome”. Science, Vol. 310, 14 October 2005, pp. 239-240. J. Furman and S. Stern (2006). Climbing Atop the Shoulders of Giants: The Impact of Institutions on Cumulative Research. NBER Working Paper.