The problem of regional imbalances in technology-based innovation is hugely important. Standard approaches to this problem tend to assume that technology-based innovation and resultant economic growth will automatically occur if
· local research institutions are sufficiently strong;
· the regional concentration of high tech firms is sufficiently high;
· the regulatory environment is sufficiently favorable to risk capital,
· political leadership is sufficiently strong.
All of the above "conditions" for regional innovation are necessary, but they are not sufficient. No amount of government support for a nascent venture capital industry will help a region devoid of technology entrepreneurs. A state-funded technology park established near a top research university will house few tenants if the university discourages faculty from seeking to commercialize their innovations. An isolated cluster of heavily subsidized start-up firms constitutes an innovation ecosystem no more than an assemblage of parrots and potted ferns is a rainforest.
Private sector actors in the innovation system, informed policy makers, and academics all know that, to paraphrase [former House speaker] Tip O’Neil about politics with modest hypebole, “all innovation is local.” However, the more specific questions that engage policy makers lie beyond our current state of knowledge:
- Where an innovation-based economy does exist, how does government act (or refrain from acting) to support its continued growth? Where one does not exist, what can be done to encourage one to develop?
- What are the critical links in the innovation network? What opportunities, if any, exist for partnerships between governmental bodies operating at different scales or in neighboring jurisdictions and various actors in the innovation system? What should be the roles of local, state, and federal governments in supporting innovation?
- In an ideal world, what programs should local, state, and federal government fund, how much should the programs receive, and how should their success be measured?
A better understanding of regional innovation, and thus better public policy and program design, depends on being able to answer the follow questions:
- What fundamental set of behaviors, contracts and incentives characterizes the “regional ecology of innovation”? By what processes are basic scientific breakthroughs translated into commercializable products and processes? What of incremental improvements to products and processes?
- In innovation ecosystems, how do networks of relationships and trust form? What are their limitations of scale and scope (e.g. geographical scope, number of names in the "Rolodex," qualitative variety of contacts)?
- How do the local, state, and federal governments engage in the innovation system, positively and negatively? What kinds of policies and programs at each scale of government best support currently thriving innovation ecosystems, and which best nurture the development of new ones?
- To what extent is regional specialization—e.g., the development of regional technology “clusters”—truly a requirement for successful competition in global markets for knowledge-based goods and services? Does regional specialization increase the likelihood of capturing economic gains from innovations locally? In an era when technologies, products, and services are increasingly developed upon shared platforms, with networks of research centers, suppliers, and customers linked in complex ways across industry boundaries, are clusters less important? How relevant today are assumed boundaries between ‘traditional’ economic activities (e.g. textiles, fishing, and agriculture) and new, technology-based industry areas considering, for example, advances in robotics and ag-biotech? How do new technologies provide p remote regions overcome traditional limitation in the digital economy, given new technologies?
We use the terms “ecology” and “ecosystem” metaphorically. At the same time, we recognize that the project may benefit from exploring further the insights for human systems of research by leading ecologists and evolutionary biologists (see e.g. Levin 1992). The use of evolutionary and ecological metaphors in economics has a long history in economics, dating back at least to
(1890). Yet natural system differ from human social systems in fundamentals respects. Marshall
In the 1950s and 1960s many economic models informing public policy assumed that basic science (and spinoffs from military R&D) “automatically” led to productivity and market growth. However most industrial innovation is based on modest extensions of existing technology, so most productivity gains result neither from advances in basic science nor from radical new technologies, but rather from steady improvements to existing innovations. During the 1980s, Japanese success in the high tech industries in which
firms had been dominant were finally understood as a failure of domestic firms to adequately prioritize manufacturing efficiency and consumer satisfaction. In the 1990s the U.S. economy surged. Growth derived fundamentally from a dramatic increase in productivity. While the sources of these productivity gains are debated among economists, most attribute a substantial part of this growth in real terms to the surge in the creation of entrepreneurial, venture capital-backed, technology firms. U.S.
At the end of the 1990s, perceptions shifted again toward a view that technology-based radical innovations uniquely replenish the economy’s long-term potential—though their short-term contribution to economic growth is relatively minor. The role of technology entrepreneurship in general, and particularly the institution of venture capital, as engines of
economic expansion became almost an article of faith among politicians, pundits, policy-makers and the public. The stampede of investors into, then out of, the public market for equity in technology-based “new economy” firms and the volume of traffic from U.S. Silicon Valley to Wall Street came increasingly to represent not just a single economic sector, but rather the scorecard for the economy as a whole.
The actual relationship between innovation and job growth and venture capital investment is a considerably murkier problem than newspaper accounts or stump speeches would suggest. Innovation today is usually the product of many different types of entities working together. The techno-wizards, perhaps joined by a few others in small start-up efforts, bring their new ideas. Venture capitalists and angels add their money, their contacts, and their vision of the marketplace. Strategic partners offer downstream or upstream customers, rapid access to established marketing networks, and swift scalability in manufacture. Investment bankers proffer the end-of-the-rainbow pots of gold. Beyond this, universities may provide basic technologies for license, governments key incentives for location, lawyers effective structures to allocate shares and align incentives. Specialization across firm types and corporate boundaries characterizes today's successful high tech firm. The success of each of these firm types, the metaphorical equivalent of species in an ecosystem, depends on the presence of others. Understanding the detailed characteristics of the processes supporting innovation at a microeconomic level is critical not only to resolving their impacts at the macroeconomic level, but to designing public policy that allow them to function efficiently and equitably.
Schumpeter (1912) foreshadowed current discussions, referring to the role of the entrepreneurs in bringing together resources to create “new combinations” of economic activity—ones that occasionally succeeded in challenging incumbent forms of economic activity. At the start of the 21st century, what do we know about the process by which technology entrepreneurs in partnership with venture capitalists, corporate technology managers, university technology licensing officers and others in the innovation system conceive and implement “new combinations”? Though our understanding is fragmentary, we can make some observations:
- Both technologies and the markets in which they are bought and sold are becoming increasingly complex. Arriving at a detailed understanding of either requires many years of painstaking effort. Yet, having reached the frontier of knowledge regarding a technological area or a market, an innovator cannot rest for long because knowledge depreciates rapidly. Markets exert pressure to standardize and modularize. Complexity on one hand and standardization on the other are thus almost yin and yang forces, the former reinforcing the traditional role of personal contacts, the latter pushing the drive toward impersonal markets.
- Nearly all new technological applications arise either from incremental change to, or new combinations of existing technologies. Understanding how to combine existing technologies (e.g., the internal combustion engine and the air-foil) to create a new product (e.g., the airplane) requires having some understanding of how each of the technologies works separately. As technologies become more complex, new combinations require collaboration.
- The key obstacle to funding a technological collaboration—bringing into existence a new technological and/or economic combination—is the ability to identify the minute subset of potential combinations for which a viable market exists and which matches the skill set of a group of specialized technologists whose services can be engaged—a task referred to by venture capitalists as “due diligence.”
- Technology entrepreneurs are not, as a rule, able to carry out this work on their own. To build one’s knowledge base to the point where one is working at or near the technological frontier and to organize a quality research team is a challenging enough job description. Most technology entrepreneurs cannot additionally maintain a venture capital caliber network of business contacts.
- Quality venture capitalists and angel investors must be able to evaluate the quality of new combinations. They must also be able to gain and maintain a good knowledge of the abilities of a large group of potential participants in such projects, particularly those with high levels of technological abilities. Furthermore, venture capital companies continue to add value along the way. They do this in partly by helping companies develop their business plans, products, and marketing strategies. They also do this extensively through their reputations and connections—making introductions, and attesting to quality. The success of the best venture capital firms (those that capture a disproportionate share of the returns in the industry) depends far less on their ability to pick winners than on their ability to create winners. This both adds value to the firms they fund and enables them to attract the most promising firms. Given the particular difficulties of the contracting for technological information, barriers to entry will be high, so individuals and firms that successfully manage such contracting should reap high rewards (a theoretical prediction readily supported by data).
- The effort to value technological information—assessing the market possibilities of new, perhaps recombined, technologies—is likely to be severely constrained; few skilled individuals are available to evaluate of new combinations. There will be always be many more potential new combinations of technologies—imaginable, but not tried—than there are companies and their financial supporters to try them out. The gap between potential breakthrough ideas and the number that receive a fair trial may grow in the future, as the execution of new technological combinations increasingly requires the collaboration of different actors with specialized skills.
- Modern technologies require the continual exchange of information and products; yet traditional markets can not accommodate such transactions. In response, contracting is now accomplished with alternative arrangements. The firmest ties come through a merger. Incentives get aligned, information exchange is no longer guarded. At the opposite extreme we see alliances, where two or more firms work together for a time, often with capabilities but not dollars changing hands. Alliances may even be institutionalized, as they are with the Internet, where no charge is imposed by any provider to carry the information initiated by others. Diagrams of firms’ relationships with strategic partners often have dozens or even hundreds of lines of connection.
Historical evidence amply documents the presence of significant knowledge spillovers and other intra-industry increasing returns to scale within regions—automobiles in Detroit, venture capital in Silicon Valley, biotechnology in the Boston metro region and carpets in North Carolina. Persuasive theoretical arguments, and some empirical evidence, support the claim that sustained regional growth requires not only the presence of specialized industry clusters, but also a certain degree of economic diversity. Barriers to entry in the field of any complementary capability will hurt the entrepreneurial industry as a whole. Not surprisingly, the most innovative regions of the country have seen rapid entry in such fields as venture capital, or technology-oriented law firms. Countries or regions with rigid regulatory structures for investment entities will find themselves disadvantaged way beyond what we would expect from a purely first order analysis.
Nonetheless, generalizing from the experience of particular regions is dangerous. Reason why include the following:
- Regions vary not only in their economic structure, but also in their culture and history. Even within regions of the U.S., attitudes towards trust, reputation and risk vary significantly.
- Ex post analyses may infer incorrectly infer causation from chance.
- For competing regions just as for competing firms, entering a market in which there already exists an established incumbent—e.g. a Silicon Valley—is very different from creating a new market.
 The National Venture Capital Association (www.nvca.org) reports that, in 1999 for example, 76 percent of venture capital investments were concentrated in four states. Within Massachusetts, for example, multiple initiatives have failed to generate economic growth based on high technology innovation in various regions outside the Boston metro. Other areas, such as the route 495 corridor, have begun to develop a regional ecology of innovation relatively spontaneously.
 A state with a limited technology base might reasonably do as West Virginia and Arizona have done in the fields of biometrics and optics, respectively: seek to nurture and support specific innovation/industry clusters from the ground up. However, in regions with a well developed innovation system such as that in the Boston metro area, there is a far greater reason to believe that a large scale, targeted state program would distort, rather than enhance, private incentives.
 Alic et al. (1992).
 Dertouzos, Lester and Solow (1989).
 Important recent contributions to this literature include Jorgenson and Stiroh (2000) and Nordhaus (2001). Bresnahan and Trajtenberg (1995) directly address measurement issues involved in assessing the contribution to growth of “general purpose technologies.”
 Such perceptions are reinforced by the astounding growth and magnitudes of venture capital disbursements—in 2000 alone exceeding $100 billion.
 Gompers and Lerner (1999, p. 137) note: “Demonstrating a causal relationship between innovation and job growth on the one hand and the presence of venture capital investment on the other is, however, a challenging empirical problem. To what extent are the mechanisms [of venture capital] uniquely suited to addressing the needs of entrepreneurial, high-technology firms? To what extent is venture capital just one of many financing alternatives for these firms, with its own set of strengths and limitations? This topic will reward creative researchers in the years to come.” A recent paper by Kortum and Lerner (2000) represents are rare, or possibly unique attempt to isolate the contribution of venture capital to innovation in the U.S.
 David Teece (1987) introduced the concept of “complementary assets” to describe these external dependencies that govern subsequent economic success of high tech innovations. This has been expanded into the notion of the economic efficiency of social capital, Fountain (1998).
 This combinatorial approach to innovation has been taken up recently by Romer (1996), Weitzman (1998; see quote above), and Auerswald, Kauffman, Lobo and Shell (2000). Romer (1996: p. 204) suggests likens combinatorial innovation to the discovery of new recipes:
New growth theorists... start by dividing the world into two fundamentally different types of productive inputs that can be called “ideas” and “things.” Ideas are nonrival goods that could be stored in a bit string. Things are rival goods with mass (or energy). With ideas and things, one can explain how economic growth works. Nonrival ideas can be used to rearrange things, for example, when one follows a recipe and transforms noxious olives into tasty and healthful olive oil. Economic growth arises from the discovery of new recipes and the transformation of things from low to high value configurations.
 The increasing complexity of both technological and market environments pressures venture capital firms to try to develop synergies through specialization within their own domain. Where one firm nurtures contacts in Internet advertising, another does so in biotechnology. The specialization goes beyond expertise to the structure of the network of relationships developed by the firm.
 Consider, for example, the PC—for some years now a mature technology. The insides of a personal computer suggest a production process easy to decentralize and distribute among a large number of fiercely competing small firms. Modular standards have been clearly established. Prices are falling precipitously. Outsourcing of production is the norm. Yet, at the same time, there are dozens of new technologies appearing on the horizon that are requiring complex contracting, networks and trust. These are the new combinations—amply associated with uncertainties, informational asymmetries and unknowables—that are well suited to the venture and angel mode of support.
 See Somay and Teece (2000) and Tassey (2001) for a further discussion of increasing technological complexity and its implications for innovation policy.
 Reputations and the link to networks explains why a new MBA hired by a leading VC firm like Kleiner Perkins can generate a million dollars of business, but that same MBA forming a new firm with three peers could not expect to generate a fraction of this business.