Michael Andrews, Aaron Chatterji, and Scott Stern have edited a compilation which examines the relationship between entrepreneurship and economic growth across sectors and its trends, the impact of policies and institutions to spur entrepreneurship on economic growth, and how entrepreneurship affects economic performance and social progress.
They highlight the contributions of Chris Forman and Avi Goldfarb, and Erik Brynjolfsson who examined innovation and entrepreneurship in the IT sector and found,
While IT has rightly been held up as the model of dynamism, Forman and Goldfarb argue that the IT sector is increasingly showing signs of becoming mature and less dynamic. With a deep dive into patent data, they show that the IT sector has become increasingly geographically concentrated in Silicon Valley, patents increasingly come from a smaller number of firms, and those firms increasingly tend to be incumbents. In his discussion, Erik Brynjolfsson reminds us that patents are an imperfect measure of innovation, and this may be particularly true for software. Nevertheless, Brynjolfsson highlights several other metrics that tell a similar story to Forman and Goldfarb. Namely, high-IT industries are more concentrated using various measures, and the often-intangible assets that are complementary to IT are increasingly found in superstar firms.
This sits in contrast to the conventional wisdom about entrepreneurial startups being at the leading edge of innovation in technology sectors.
Julian Alston and Phil Pardey surveyed technologies in agriculture.
Alston and Pardey survey the many labor-saving technologies in agriculture over the last 75 years, and consequently the dramatic decline in labor working in agriculture, the small decline in land used for agriculture, and the increase in agricultural inputs (e.g., pesticides and herbicides) and capital (farm machinery)... possible to construct detailed estimates of return to R&D. These estimated returns are massive, with estimated median internal rates of return ranging across studies from 12-41% per year, and benefit-cost ratios ranging from 7 to 12. Notably, these estimated returns are calculated over many years, and it can take decades for R&D to manifest itself in the productivity statistics. Alston and Pardey also review adoption lags for numerous agricultural technologies, and likewise find 30-50 years between when a technology is introduced and when it is widely adopted... The authors also analyze numerous more recent technologies, including precision agriculture, variable rate seeding and fertilizer, the use of satellite imaging, auto-steering on tractors, and more, and find much slower adoption rates. In his discussion, included in this volume, Brian Wright... speculates that, over time, farmers have realized that they appropriate a relatively small share of the returns to public research in agriculture, and have instead turned their attention to lobbying for market distorting policies that favor their interests.
An area of high growth highlighted is "supply chain traded services" - or services sold to businesses or government in the process of producing a separate final product. This area has benefited enormously from ICT by way of making many services tradeable and also scalable.
They point to skews in public expenditures on R&D,
In terms of the innovation funded by the federal government, more than 40% of R&D dollars go to the Department of Defense, 27% go to Health and Human Services, and 12% goes to the Department of Energy. That leaves only about 10% of federal R&D to go to all other programs, including NASA, the NSF, agricultural research, etc... What is striking is how little research is done in areas such as education, housing, and the social sciences—not just as a share of the overall federal research budget, but in absolute terms as well—despite the fact that these areas are major federal policies... One of these sectors that has received massive amounts of research spending from both the federal government and private sources is the health sector. But this research tends to overwhelmingly be directed towards new drugs, with a relatively small share of research directed towards health services. Amitabh Chandra, Cirrus Foroughi, and Lauren Molstrom investigate the health sector, with a particular focus on venture capital-led entrepreneurship, and report that 60% of VC investment in health is directed towards firms working on pharmaceuticals, 20% to firms working on medical devices, and only 20% to firms working on all aspects of healthcare delivery and infrastructure... the NIH allocates a larger share of funding towards pharmaceuticals, and less towards health care delivery, than does the private sector. Together, these facts raise the possibility that public funding is not working to resolve market failures in the health care sector.
There may be a case to be made out for crowding-out of public R&D towards the more tangible, representative, and commercially attractive parts of the innovation space and away from the purely public goods side.
Similar skew exists in housing market innovation,
While there has been little change in how housing units are constructed, there has been the appearance of numerous real estate technology firms, either tools to use the internet for housing search like Zillow or online home sharing platforms like AirBnB. While these new firms do not increase the productivity of housing construction, they do increase the match quality between home buyers and sellers, and Kung argues that this can represent substantial gains to consumer surplus... Land use regulation in particular can stifle the supply of new housing and depress incentives to innovate in the sector; Hseih and Moretti, for instance, conclude that land use restrictions could have reduced the GDP growth rate by as much as one third... The solution is an expansion in the housing supply, but both Kung and Handbury note innovation in the production of the housing stock is unlikely without policy reforms such as a reform of the aforementioned zoning and land use regulations.
Andrews and Co try to summarise these findings by drawing the distinction between technology and product segments and the non-technology and service delivery segments,
To a first order, the sectors that have seen successful innovation and entrepreneurship have been science-based (the “productivity drivers,” in particular IT, energy, and agriculture) or have been able to incorporate technologies from those fields (manufacturing and the “on- demand” sectors). In the sectors for which progress has been more mixed, such as health care, the parts of the sector that rely on science have typically seen large advances (i.e., pharmaceuticals and medical devices), whereas those that do not have largely stagnated (health care delivery, financing, non-pharmaceutical health interventions)... we find it telling that the areas that the innovative and entrepreneurial sectors are those for which an innovation system is well established. By innovation system, we mean not only well-funded public institutions to conduct R&D, although such an institution is certainly in place for the innovative sectors (i.e., the NIH, NSF, numerous large R&D projects funded by the Department of Defense and Department of Energy), but well- defined research jobs, career ladders, rewards for innovative success, and an ecosystem in place to develop and support high growth entrepreneurs. We term these sectors for which an established innovation system is in place the “Vannevar Bush sectors.”... the evidence leads us to suspect that constructing innovation systems for the non-Vannevar Bush sectors will lead, after a long delay, to technological and entrepreneurial opportunities in these areas... Many of the non-Vannevar Bush sectors are focused in the social sciences, and more specifically in determining how to innovate complicated systems with many stakeholders.
The important point is that innovation and entrepreneurship appears confined to areas involving technology, and related products and solutions. Areas involving access, affordability, service delivery quality, financing and so on remain deeply under-researched and starved off innovation and productivity improvements. Business models, delivery models, process re-engineering, organisational changes, skill development, and so on are the areas most relevant to the latter.
No comments:
Post a Comment