Brookings: Series - Issues in Technology Innovation

Becoming a Techno-Industrial Power: Chinese Science and Technology Policy

Joel R. Campbell — Mon, 29 Apr 2013 09:05:00 -0400

Editor's Note: This paper by Joel R. Campbell, which outlines the history of Chinese science and technology innovation since the founding of the People's Republic, is the April 2013 installment in the Issues in Technology Innovation paper series, which is part of the Center for Technology Innovation in Governance Studies at Brookings.

China’s science and technology policy has developed through four phases since the founding of the People’s Republic in 1949. In the first phase, to 1959, technology supported the creation of heavy industry along Soviet lines, while the second, up through the end of the Cultural Revolution in 1976, saw economic stagnation and ideological domination of technology projects. A third phase, under reforms launched by Deng Xiaoping and carried forward by Jiang Zemin to 2001, stressed building of an independent research base and the gradual shift to market-oriented, product-driven research. Since 2002, Chinese policy has increasingly backed high technology industrialization, along with support for the nascent green technology industry.

Chinese technology policymakers also have promoted an innovation-driven economy. The Ministry of Science and Technology (MOST) is the key policymaking and policy coordination organ, and it funds the five most important technology development projects:

The Key Technologies Research and Development Program, focused on industrial technology
The 863 Program, centered on basic and applied research on marketable technologies
The Torch Program, which supports commercialization of high tech products
The 973 Program, funding multi-disciplinary projects in “cutting edge” technology, and
The Spark Program, promoting development and use of technology in rural areas

Science and industrial parks are key venues for high tech research and development (R&D). Currently, there are fifty-four such parks, mostly located in large cities or provincial capitals. Firms operating in the parks must create or apply technology in high tech fields, devote at least three percent of gross revenues to R&D, and employ at least thirty percent of college degreed workers. The information technology (IT) industry is one of the leading industries in the science parks, and has received special policy recognition since 2000.

The space program has become one of China’s proudest recent accomplishments. Building steadily on its experience with military and civilian missile technology, China has already launched four manned space missions, and has ambitious plans for a space station and unmanned exploration of the Moon, along with possible manned lunar missions. China has also made a major push into green (or “clean”) technology, driven by twin concerns about dependence on foreign oil and serious environmental degradation within China.

Download the paper » (PDF)

Downloads

Download the paper

Authors

Joel R. Campbell

Image Source: © Jason Lee / Reuters

Invention and the Mobile Economy

Darrell M. West — Tue, 05 Mar 2013 15:00:00 -0500

In this paper released in conjunction with a Mobile Economy Project panel discussion, Darrell West argues the importance of invention to mobile communications and demonstrates that the mobile industry is one of our most vibrant drivers of economic development.

Highlights include:

• Examples of key inventors: West seeks to understand how to sustain invention and draws lessons for encouraging the critical innovation needed for future development.

• How different countries handle invention: In comparing and contrasting other countries as well as the United States, West adds perspective and paints a global invention landscape.

• Ways to facilitate invention: A number of factors affect the quantity and quality of invention (including but not limited to research and development, the quality of STEM (Science, Technology, Engineering and Mathematics) education, the nature of immigration, and the patent system. West emphasizes how we should maintain a culture of invention to encourage future prosperity.

Download Paper » (PDF)

Downloads

Download the paper

Authors

Darrell M. West

The Internet, Cross-Border Data Flows and International Trade

Joshua Meltzer — Mon, 25 Feb 2013 00:00:00 -0500

The Internet is becoming a key platform for commerce that is increasingly happening between buyers and sellers located in different countries, thereby driving international trade. Additionally, as the Internet enables cross-border data flows this is also underpinning global economic integration and international trade. For instance, cross-border data flows are now intrinsic to commerce, from Internet-based communications like email and platforms such as eBay and Facebook that bring buyers and sellers together, from the financial transaction to purchase the product in other countries to the downloading of the goods and services.

Despite the growing significance of the Internet for international trade, governments are restricting the Internet in ways that reduce the ability of businesses and entrepreneurs to use the Internet as a place for international commerce and limits the access of consumers to goods and services. Some of these restrictions are being used to achieve legitimate goals such as preventing cybercrime or restricting access to morally offensive content, but may be applied more broadly than necessary to achieve those objectives. In other cases, Internet restrictions are targeting foreign businesses and the sale of goods and services online in order to benefit local ones. Such Internet restrictions are discriminatory and harm international trade.

In this paper I discuss the importance of the Internet and cross-border data flows for international trade. I propose steps that governments should take to apply existing international trade rules and norms and identify where new trade rules are requires to further support the Internet and cross-border data flows as drivers of international commerce and trade.

Building on the international trade law developments I describe in the paper, the following outlines the key challenges that remain and proposes ways trade policy and law could address them:

Develop binding commitments with exceptions: trade rules should establish cross-border data flows as a mandatory legal norm while providing sufficient policy space for governments to restrict data flows where necessary to achieve other legitimate policy goals. Such restrictions should also be designed and applied in a non-discriminatory, least trade restrictive and transparent manner.
Intra-Country Data Flows: commitments on cross-border data flows should include a commitment to also not restrict intra-country data flows. There is no commercially sound reason for rules on cross-border data flows to not also apply to their movement within a country. And once data is allowed across-borders, many of the reasons for government restrictions on intra-country data flows diminish if not entirely disappear.
International standards: global industry standards and interoperability criteria will underpin growth in cross-border data flows, such as the ability of users to access and use digital content across devices. Governments should commit to developing international standards with the aim of underpinning technology development that is consistent with Internet operability.
Location of Data Centers: requiring data center to be located domestically undermines the cost-effectiveness of cloud-based computing services where so-called location independence is important. Under KORUS the parties have addressed this issue for the financial sector by agreeing to allow financial institutions to transfer data across their borders for data processing. Governments should commit to similar rules for all cloud-based computing services.
Rules on transparency: Internet restrictions on cross-border data flows are often implemented in an arbitrary and non-transparent manner. Some FTAs have sophisticated rules requiring transparency and due process, but this is yet the norm. Moreover, Internet restrictions on cross-border data flows raise specific issues that require additional commitments in the following areas:
- A designated contact point in the government agency responsible for restrictions on cross-border information flows.
- Provision of advanced notice of any proposed measures affecting cross-border data flows, including the reasons for the proposed restriction.
- Opportunities for interested parties such as businesses or individuals to present their views on the proposed restriction and a requirement for written and reasoned responses.
- Opportunities for administrative review of Internet restrictions.
Develop Norms on Cross-Border Data Flow: governments should also prioritize developing norms of conduct amongst governments with respect to the Internet. In addition to the role of binding trade rules here, governments should develop principles governing access to and use of the Internet. For example, the US and Japan have agreed to Internet principles that emphasize the preservation of an open and interoperable Internet and a balanced approach to issues such as privacy and intellectual property rights so as not to impede the cross-border flow of information.
Address the digital divide: For businesses in developing countries, non-tariff costs such as inadequate logistics and transportation services have a significant impact on the costs of exporting. As noted, increasing Internet access in developing countries can reduce costs of exporting by up to 65 percent. Assisting developing countries better integrate into the global trading system should therefore include increasing Internet access and the provision of cheaper mobile devices to access the Internet.

Downloads

Download paper

Authors

Joshua Meltzer

Image Source: © Thierry Roge / Reuters

Technology Optimism or Pessimism: How Trust in Science Shapes Policy Attitudes toward Genomic Science

Jennifer Hochschild, Alex Crabill and Maya Sen — Thu, 03 Jan 2013 00:00:00 -0500

Just as physics gained public visibility and ideological contention as it matured over the twentieth century, so genomic science will gain public visibility and competing normative valences as it becomes increasingly important during the twenty-first century. The new biology can help to solve many of humankind’s most serious problems, or it can reinforce racial hierarchy and enhance governmental surveillance. Genomic scientists will protect our lives and our planet -- or scientists are tampering dangerously with nature.

As of 2013, the production and use of genomics has been lightly regulated; court cases addressing medical or individual uses of genomics are rare and politicians’ engagement even rarer. The American public is just beginning to learn about genomic science, its likely uses, and its potential benefits and harms. This relative vacuum permits social scientists to explore how innovations in genomics research are moving into the public arena; rarely do scholars have the chance to watch a new policy regime emerge, especially in such an important and fraught field. As public opinion develops, it may help to shape government funding, regulation, and legislation. Most importantly, to the degree that biology becomes in this century what physics was in the last – a powerful, somewhat mysterious force influencing the destinies of individuals, countries, and the globe – it is essential for any democratic polity to examine what people know, want, believe, and fear about it. Those views will surely change as the science changes, but baseline analyses will enable political actors to know where people are starting.

We begin by describing Americans’ level of technology optimism or pessimism across four arenas in genomic science and one arena (climate change) outside genomic science. We then ask “so what?” – do people who perceive more harm than good in genomic science hold different policy preferences from those who perceive more good than harm? Do optimists and pessimists differ in their perceptions of elite actors, or their willingness to be directly involved with the new science? Finally, we consider variations within the public. Is knowledge about genetics associated with more optimism about genomic science? Are people with direct interests in one arena of genomics more optimistic (or pessimistic) about its future than they are about other arenas? Do religiosity or characteristics such as race or gender play a role in levels of optimism about genomics in general or particular genomics arenas?

We conclude, first, that public attitudes toward genomic science are coherent and intelligible, perhaps surprisingly so given how new and complex the substantive issues are, and, second, that citizens differ from most social scientists, legal scholars, and policy advocates in their overall embrace of genomics’ possibilities for benefitting society.

Download » (PDF)

Downloads

Download Paper

Authors

Jennifer Hochschild
Alex Crabill
Maya Sen

Image Source: © Jason Reed / Reuters

Improving University Technology Transfer and Commercialization

Darrell M. West — Wed, 05 Dec 2012 00:00:00 -0500

The federal government invests $147 billion in U.S. research and development, with $90 billion going to institutions of higher learning to underwrite faculty research projects and the training of graduate students and post-doctoral fellows. However, based on licensing fees, federal dollars generate a very small rate of return on investment.

In this new paper, Darrell West examines that low ROI and argues that part of the problem is that the focus on patents, licenses, and startups places too much emphasis on outputs as opposed to outcomes. Those indicators represent proxy measures of getting material to the market as opposed to whether particular research ideas actually are having an impact and being successful in the marketplace. If a patent is awarded, a license issued, or a start-up business established, it does not guarantee that the product is used or generates revenue.

This paper reviews how universities report their commercialization activities, the need for better performance metrics, and ways to improve their disclosures and overall performance. Using an analysis of technology transfer annual reports, West argues that universities should provide more detailed financial performance data. By offering more complete material on money in and out, it would help evaluate how well universities are commercializing their research ideas and whether alternative models would produce better results. There needs to be better understanding of the innovation differences across academic fields, and increased emphasis on university transparency, accountability, and overall performance. West gives specific recommendations for ways to do better on technology transfer and commercialization, including standardized reporting forms, equity investments and greater transparency.

Download » (PDF)

Downloads

Improving University Technology Transfer and Commercialization

Authors

Darrell M. West

Image Source: © Norbert von der Groeben / Reuters

Building an IT Economy: South Korean Science and Technology Policy

Joel Campbell — Wed, 12 Sep 2012 00:00:00 -0400

South Korea has created a robust IT economy and in this paper, Joel Campbell explores how the country's recent post-war history has encouraged this technology and science development.

EXECUTIVE SUMMARY

Underlying Korea’s remarkable post-1961 economic development has been the development of a strong science and technology capacity. During the authoritarian regimes (1961-1988), the state created a rudimentary research capacity, primarily focused on creation of government-run research institutions, a technical university, and a central research park, as the private sector gradually began to muster its own applied research capacity. The late 1980s to late-1990s saw a change of direction, as Korea’s chaebol conglomerates became the lead actors in R&D. The well-funded National S&T Technology Program became the focus of state efforts, later superseded by the 21st Century Frontier Program and specified research funds. By the turn of the century, Korea had achieved strong aggregate performance in terms of numbers of researchers and funds spent on R&D, and has continued to build on that for the past decade.

The IT industry and, to a lesser extent, biotech have become the major drivers of technological development. The shift from the old industrial to new high tech economy facilitated a recasting of national efforts. A refocused state helped midwife the nascent IT sector, through a combination of privatization of the national telephone service provider, creation of infrastructure, and dispute moderation. Even so, recent doubts about Korea’s overall IT competitiveness have arisen.

Meanwhile, since the mid-1990s, Korean policymakers have been captivated by the possibilities of “Big Science,” i.e., basic or foundational science. Korea participates in various international basic science programs, and has created another big state funding effort (the 577 program) to support basic science. The government has spent much policy effort on drafting “visions” of future technological development, but its technological future may depend on maintenance of economic competitiveness.

The fast post-war development of South Korea is one of the most remarkable economic stories of the twentieth century. The small Asian nation in 1960 was one of the world’s poorest countries, with a Gross Domestic Product roughly equal to that of Ghana. By 1995, it rose to become the twelfth largest economy, and Asia’s fourth largest. How Korea was able to accomplish this remarkable feat is a much analyzed story in international political economy, but at its heart was a largely autonomous state that employed a combination of state-directed bank financing, light and then heavy industrial export promotion, fostering of large industrial conglomerates (the fabled chaebol), and suppression of labor unions to create workplace peace. A hard-line military regime gave way to democracy from the late 1980s onward, and the state committed to thoroughgoing economic liberalization as a result of the calamitous 1997-1998 Asian Financial Crisis. Since the late 1990s, Korea has generally been accepted as an addition to the developed world and, as a mature economy, has seen economic growth slow and population size plateau. Koreans have become intense users of electronic media, with broadband computer connectivity and cell phone service achieving nearly universal penetration.

Underlying Korea’s strong economic development has been a consistent effort to create a robust science and technology (S&T) capacity. From the beginning of Korea’s export-oriented drive in the 1960s, this has followed two parallel tracks: creation of a state-led research and educational capacity, centered on state-run research institutes, and in-house research and development efforts by the chaebol and some medium-sized firms. Universities were a relatively weak S&T player, at least until the late 1990s. After the mid-1990s, the focus of state S&T policy shifted from industrial technology to promotion of the information technology (IT) industry.

To get a better understanding of Korea’s technological development, this article examines the post-1961 history of technology development, and the transition to an IT-dominated economy in the 1990s. It then examines state policy and institutional changes, and promotion of the technologies of the twenty-first century. It considers state policy in global science and technology, what Korean technology writers call “Big Science,” and Korea’s future as a technology power.

Download » (PDF)

Downloads

Building an IT Economy: South Korean Science and Technology Policy

Authors

Joel Campbell

Image Source: © Kim Hong-Ji / Reuters

Digital Music Broadcast Royalties: The Case for a Level Playing Field

John Villasenor — Tue, 07 Aug 2012 12:15:00 -0400

Royalty rates play a crucial role in shaping the digital music broadcasting industry. If rates are too high, the ability of digital broadcasters to provide the public with access to music is impeded. If rates are too low, then recording artists do not receive a fair return on their endeavors. And if rates are inconsistent across different delivery mechanisms, then some business models are favored over others.

Today, the playing field in American digital music broadcasting is anything but level. For example, Internet radio companies can be compelled to pay over 60% of their revenue in sound recording performance royalties. By contrast, Sirius XM satellite radio currently pays only 8% of gross revenue. To make matters even more complicated, these rates are evolving over time in complex and sometimes unpredictable ways, making it nearly impossible for digital audio broadcasters to make reliable projections regarding their future content acquisition costs.

This new paper examines how the current patchwork of rates reflects a two-tiered system in American copyright law that partitions non-interactive digital audio services into two categories, each with its own standard for statutory performance royalty rate determinations. Services such as Internet radio that have the misfortune of being subject to the more onerous “willing buyer/willing seller” standard can face extremely high rates, while those such as Sirius XM satellite radio that are often associated with the more balanced standard known as “801(b)” enjoy much lower rates. Furthermore, access to the 801(b) standard is limited to certain services that were “preexisting” in 1998, the year that the Digital Millennium Copyright Act was enacted.

The current copyright royalty landscape creates significant inequities among current market participants. It also furnishes a strong disincentive to potential new market entrants and to the introduction of innovative new business models for delivering digital music. The good news is that these shortcomings can be addressed through simple, focused legislation to provide all non-interactive digital music broadcasters – not just a favored few – with access to statutory royalty rates determined according to the 801(b) standard. This would stimulate innovation and growth in a key segment of the broadcasting industry.

Download » (PDF)

Downloads

Digital Music Broadcast Royalties: The Case for a Level Playing Field

Authors

John Villasenor

Image Source: Steve Marcus / Reuters

Communication Technologies: Five Myths and Five Lessons from History

J. P. Singh — Mon, 02 Apr 2012 00:00:00 -0400

EXECUTIVE SUMMARY

Mobile phones in the developing world have myriad uses: banking services, reminders for medicine regimens, e-governance, and more. This is a far cry from a generation ago when 99 percent of the people in low-income countries lacked POTS, or “plain old telephone service.”

Information and communications technologies are now indispensible for development, prioritized through varying levels of market-driven measures and participatory politics. From international organizations to local administrations, the importance given to these technologies for development today is a counterpoint to the immediate post-colonial era when telephones were considered a luxury and nationalized radio broadcasting was used for bringing “modern” ideas to populations. Along with policy changes, the move toward market forms works to ensure that people have phones and access to communication infrastructures, in turn providing incentives for entrepreneurs and political brokers to develop applications for delivery of social services and provide alternatives to users who in an earlier era lacked even basic access to these technologies.

Information technology diffusion rates can be quite spectacular. Only one in a thousand people had a mobile phone in 1995 in low-income countries. Now more than 25 in a thousand people do. Social media and the Internet have revolutionized political participation globally and provided voice and solidarity to communities. In January 2008, a 33-year-old civil engineer from Bogotá used a Facebook page to organize a protest in 40 countries against the paramilitary group FARC, gathering over 12 million people. The digital divide is not fully bridged, but the exponential growth rates of political voice and telephony promise a bright future.

What lessons can policymakers learn from the last 60 years of deploying communication technologies for development? Looking beyond the growth rate numbers suggests processes that either need to be continued or encouraged, but also fine-tuned at micro levels to address demands.

Encouraging Markets: Ensure regulatory independence and market incentives for providing access to infrastructures. Problems remain with corruption among officials and private firms, which calls for independence of regulatory agencies and dispute resolution, as well as smart policies to incentivize delivery in underserved areas.
Developing Polycentric Policymaking: Direct, top-down development interventions do not work effectively. International civil-society and international governmental organizations are best served as knowledge brokers and facilitators of information exchanges.
Allowing Participation and Voice: Locate political spaces for participation and voice. Development interventions tend to be expertise driven and top-down; however, it is not difficult to provide synergies between development aspirations and local contexts.
Understanding Representation: Allow people to represent themselves over various forms of audio-visual media. Old paternalistic habits are still too controlling, even as new social media defy this logic.
Prompting Ingenuity: Encourage technological and business entrepreneurship that enables political voices, social services delivery and micro-level efforts.

Downloads

Download the paper

Authors

J. P. Singh

MeTube: Political Advertising, Election Campaigns, and YouTube

Rob Salmond — Tue, 13 Mar 2012 00:00:00 -0400

YouTube provides a new channel for campaigns to broadcast their messages to citizens. How is it affecting the nature of campaigning? This paper examines YouTube videos uploaded in various campaigns in 12 countries, examining the tone, content, and popularity of each video.

Highlights include:

Tone: YouTube campaign videos are more positive than ads aired on television, but that the size of the difference between the two mediums varied greatly between countries. Informing and inspiring supporters is a task well suited to YouTube videos. Attacking an opponent, however, is more effectively done on TV. This bifurcation between a sunny YouTube presence and a mean-spirited television ad campaign is stronger in US-style winner-take-all elections than in European-style proportional elections, and has major consequences for the character of campaigning and how candidates are seen by voters.

Audience: The audience for YouTube advertisements is younger, richer, more educated, more politically interested, and more partisan than the population at large. In other countries, the differences are likely even more stark, because the U.S. has a much higher reported use of the internet for political purposes than do most other democracies.

Broadcast cost: Running long-form political infomercials on television is not unheard of, but expensive. With YouTube, it is now possible to provide detailed video information to activist copartisans and other members of the political elite for a broadcast cost of $0. The average length of each YouTube video, even including all the uploaded TV ads, is more than five minutes both across all the countries and within the U.S. Several hundred of the videos are over 10 minutes in length, often consisting of extended clips from candidate speeches. TV ads, of course, are far shorter. YouTube provides a substantially more cost effective means of providing detailed information to those who seek it than does TV.

Election Rules: Information from the observed YouTube campaigning is clear on this question: the more proportional electoral systems tend to induce more positive election advertising, while the American- or British-style single member district elections tend to be more negative. As before, this difference is statistically significant and robust to multilevel, multivariate analysis. Almost one in four YouTube videos in American-style election systems are attack ads, compared with one in six videos in proportional representation systems.

Download the paper »

Downloads

Download the Paper

Authors

Rob Salmond

Image Source: © Eric Gaillard / Reuters

Facebook, iPhones: How Evolving Mobile Technology Shapes Campaigns

Darrell M. West — Tue, 14 Feb 2012 00:00:00 -0500

EXECUTIVE SUMMARY

In 2008, candidate Barack Obama pioneered several innovative applications of digital technology. With the help of the Internet, he raised $750 million. He made use of social media platforms such as Facebook and MySpace to identify and communicate with supporters around the country. And through Meetup.com, he launched virtual get-togethers with voters in many different locales simultaneously.

Four years later, we see another wave of innovation enabled by mobile technology in the United States and around the world. Smartphones and handheld devices have proliferated and now outnumber desktop computers. Candidates, voters, activists, and reporters are using these vehicles for public outreach, fundraising, field organization, political persuasion, media coverage, and government accountability. Unlike 2008, where text messaging was the dominant feature of mobile campaign outreach, this year there has been a proliferation of mobile ads, video, web links, and apps.

As part of our Mobile Economy Project, I review innovative examples of campaign outreach made possible through mobile technology. I show how smartphones expand the opportunities for mobilization and ways in which certain policy steps would expand citizen participation.

Download full paper » (PDF)

Downloads

Download Full Paper

Authors

Darrell M. West

Image Source: Jason Reed / Reuters

Government-wide Information Sharing for Democratic Accountability

J.H. Snider — Thu, 08 Dec 2011 15:43:00 -0500

In a representative democracy, average citizens should be able to easily monitor the public actions of their representatives and the politically powerful who seek to influence those representatives. New semantic web technologies make such monitoring more cost effective to do than ever before. But while these technologies have been widely used to monitor the weak, they have not been used to monitor the powerful, who often cite privacy and cost concerns as excuses to avoid such monitoring. This paper recommends asking the president of the United States to 1) use the new technologies so the American people can more easily monitor his public, official actions, and 2) serve as a showcase for Congress and the rest of the executive branch of government-wide information sharing for democratic accountability.

Introduction

If public officials can easily track citizens (aka “potential terrorists”) across thousands of government databases, why cannot citizens do the same for public officials (aka “democratic representatives”) and the powerful political players that seek to influence them?

Tracking the Politically Weak Vs. Powerful

Consider a public official with a security clearance who wants to do a background check on a citizen. He can enter the citizen’s name in a simple query and find any suspicious activity and relationships for that citizen gathered government-wide, including by tens of thousands of local police, fire, and transportation departments. The result of this government-wide information sharing is that the public official can “connect the dots” and find patterns of suspicious behavior otherwise undetectable.

Now consider an investigative reporter who wants to investigate the official and nominally public actions of elected representatives and those who seek to influence those representatives. To do his job, the reporter, like the public official, needs to search across thousands of government databases, albeit for different information. For example, to track the official actions of a member of Congress, including possible inappropriate influences on and beneficiaries of those actions, the list of databases is long, including Congressional campaign contributions, lobbying, gifts, travel, office expenses, correspondence to Federal agencies, floor and committee votes, floor and committee remarks, press releases, bill text and sponsorships, reservations for Capitol Hill meeting rooms on behalf of private organizations, personal finances, regulatory appointments, and employment pre- and post-Congress. Federal agency databases, including agency contracts, awards, permits, leases, and licenses; comments in regulatory proceedings; advisory committee appointments; and various ethics disclosures, may also prove insightful.

The information categories listed above are conceptual, not physical. Physically, they may each be subdivided into many separate databases. For example, the U.S. House and U.S. Senate keep physically separate records; each of the dozens of Federal agencies keeps its own Congressional correspondence; and licenses/leases to use oil, timber, grazing, spectrum, building, and other public assets are physically scattered across hundreds of different Federal databases.

If an investigative reporter could search government-wide for the names of individual public officials with the same ease that public officials can search for potential terrorists, the reporter would undoubtedly be able to do his job much more efficiently and effectively.

The Semantic Web Breakthrough

Emerging so-called “semantic web” technologies make searching across government databases and websites easier than ever before imaginable. A vivid simple example is product ratings. Google created a simple set of product rating metadata (metadata is data about data) for any website to tag its product ratings for easier search. Now anyone who enters a product in a Google search can find all the ratings for that product scattered over the hundreds of millions of websites that Google scans.

Semantic web technologies also promote relationship mapping, as illustrated by Facebook’s social metadata, which makes it possible to track friendships and other types of relationships among individuals. Tracking relationships for democratic accountability is especially important because embarrassing political influence is typically laundered via intermediaries.

If government databases tagged the name of public officials with a unique identifier, the same type of decentralized search and relationship mapping could be used to track the official actions of those officials.

NIEM and UCORE

To a remarkable degree, the Federal government’s National Information Exchange Model (NIEM) already makes feasible highly ambitious government-wide information search. Launched after 9/11 to facilitate government-wide sharing of information about potential terrorists, NIEM has greatly expanded its sphere of sharing in the last few years.

NIEM creates standardized metadata that make it possible for different government databases to easily share information. The various sets of metadata may be called ontologies, and the ontologies are organized hierarchically. At the top of the hierarchy are ontologies shared by particular domains of knowledge such as the CIA and Department of Justice. As of November 2011, NIEM had twelve domains: Biometrics; Chemical, Biological, Radiological, Nuclear (CBRN); Children, Youth, and Family Services; Cyber; Emergency Management; Immigration; Infrastructure Protection; Intelligence; International Trade; Justice; Maritime; Screening; and Human Services.

At the bottom of the hierarchy is a core set of ontologies describing attributes such as who, what, when, and where that are shared by virtually all exchanges of data. A popular version of this NIEM core is called Universal Core (UCORE).[1]

Despite NIEM’s mission to facilitate information sharing, all its initial domains of knowledge were associated with secrecy, notably national security and criminal justice. Its newer domains, such as healthcare and family services, have also predominantly shared information only with a privileged set of users. Open government, in contrast, is based on the democratic norm that government information is equally available to all citizens.

However, NIEM does have the potential to be used for open government applications. For example, it was used to implement Recovery.gov, which tracks Federal contract expenditures. The reason NIEM remains strikingly weak in the area of open government probably has less to do with technology than politics.

The Core of the Core

From the perspective of open government, Who metadata may be viewed as the core of the core. This is because Who metadata is essential to holding public officials publicly accountable. If you cannot distinguish between the thousands of John Does with the same name and different permutations of the same name in government databases, government-wide search isn’t very useful.

So why not apply a Who ontology to public officials and the powerful political players who seek to influence them? The standard response, other than the increasingly implausible “it is too expensive to do,” is that this would violate the privacy of U.S. citizens. The last thing we want to do is make it easy for the U.S. government to track all the government interactions of its 300+ million citizens. This hearkens back to the image of the Federal government’s discredited Total Information Awareness program, which was unpopular even among the most ardent advocates of the Patriot Act.

But this argument turns out to be merely a convenient straw man. Not all applications of a Who ontology for open government are controversial. Indeed, some would be very, very popular among the American public. If not for the open government community’s insular, short-term, copycat, and advocacy oriented culture, which has resulted in a striking lack of public policy creativity (a controversial assessment given the many highly promoted open government innovations in recent years), it would be hard to imagine how this straw man argument could have remained unchallenged for so long.

Technically Easy But Politically Hard: A Who Ontology for Congress

A simple, cost effective, and publicly popular application of a Who ontology would be to track the official actions of the 535 members of Congress. As part of the Federal government’s Personal Identity Verification (PIV) system, more than 5 million Federal employees and contractors already have unique personal identifiers, including a unique numeric identifier linked to a unique biometric identifier (such as a fingerprint or iris scan). These identifiers are encoded on a smartcard that Federal employees and contractors must increasingly use to access Federal facilities and computer systems. The long-term goal is to prevent any access to Federal facilities or computer systems without the use of such a device. States, localities, and businesses have also started to implement the global unique ID (GUID) incorporated in the Federal government’s PIV system. Another unique personal identifier for members of Congress, such as the bioguide URL for each member of Congress issued by the Clerk of the U.S. House of Representatives, could be used for this purpose.[2]

The metadata describing the unique identifier associated with each member of Congress would be attached to every official Congressional action. For example, when a member sent an official and thus nominally public letter to any government agency, the unique ID would be attached as metadata to his or her electronic signature on the letter. Each letter could also include what, when, and where metadata. What would specify a letter, When a date stamp, and Where an address (both from and to).

A Who-What-When-Where ontology may include many standardized subcategories of metadata. For example, in addition to a postal address, Where could include an email address, a website, and a set of GPS coordinates. Similarly, When could include a point in time or a period in time.

A journalist would then be able to enter a simple query, say, to track all the correspondence of a member of Congress to the 50+ federal agencies during a particular period of time.

The problem with this exceedingly simple yet powerful Congressional application is clearly not technology, cost, or public appeal. The problem is that most members of Congress would never agree to it, except in response to an unlikely popular uprising, because information is power and they wouldn’t want potential challengers and other troublemakers to have easy access to such information.

More Politically Feasible: A Who Ontology for the President

I propose a different initial application of a Who ontology for democratic accountability, one which is the ultimate combination of simplicity and power. The proposal avoids the complexity and political difficulties of mandating the use of unique IDs for hundreds of millions of Americans or millions of incorporated organizations. It doesn’t even bother with the relatively small number of members of Congress. Instead, it initially seeks to implement a Who ontology for only a single person: the president of the U.S. The reason is that winning the president’s assent to attaching standardized who-what-when-where metadata to his official public actions is probably a lot more politically realistic than getting Congress to do the same.

President Obama has demonstrated great willingness to endorse open government practices. This proposal is a natural extension of those policies and one he could easily implement with minimal need to cajole the bureaucracy or seek any additional Congressional appropriation. Moreover, this pathetically simple and affordable technology would provide the public with a qualitatively improved tool to track his official and nominally public actions on their behalf. To assure accountability, Federal entities that receive the president’s official correspondence, including the embedded who-what-when-where metadata, would be required to post it on their public websites, although it could be buried almost anywhere on their websites because the public would primarily access the correspondence via search engines.

Once this simple application demonstrated the utility of the Who-What-When-Where ontology, it could gradually be expanded in ways the public would not only find uncontroversial but would enthusiastically endorse. One way to think of this expansion is in terms of the core of the core: the number of government officials who could be tracked in this way. For example, it could be expanded to all senior White House officials and heads of the various departments that constitute the president’s cabinet. The Plum book, published by Congress, which contains the list of the most senior positions in the Federal government, over 7,000 political appointees, would be a good target universe.

Since all Federal GUIDs in the PIV system are attached to unique identifiers for Federal agencies, if you know a Federal employee’s GUID, you also know the agency and sub-agency associated with the employee. The agency identifiers are derived from the codes used by the U.S. Department of the Treasury for structuring budget data by agency and sub-agency.

Perhaps the president’s actions would even embarrass the heads of the U.S. House and Senate into copying him, with pressure building for similar practices to trickle down to Congressional committee chairs and ultimately even rank-and-file members of Congress and their staffs.

Another type of expansion involves the type of information exchanged. Exchanging correspondence is an especially easy case. Complications arise when the president submits data to an external database not under his direct control, such as filings with the Federal Election Commission or the U.S. Office of Government Ethics. The solution would be to have the president simultaneously transmit his information to both the external database and an internal White House database published to the web. The White House database could duplicate the information on the external database and then add the new metadata, or it could only add the new metadata and use it to point to the relevant information in the external database. An analogy for this type of dual reporting is a retailer who enters sales data once but automatically submits it separately to state authorities (for paying sales tax) and its own accounting systems. Although less than ideal, the White House’s dual information reporting system would set an example for other Federal entities and encourage them to design their databases for efficient government-wide information sharing.

Another type of expansion would be to include in addition to who-what-when-where metadata the metadata associated with a particular domain of knowledge. This, as we have seen, reflects the hierarchical structure of NIEM. Such domains of knowledge tend to be orders of magnitude more complex than the simple who-what-when-where information focused on in this paper. But politically, when it comes to open government domains, there may be fewer obstacles to their implementation. This may help explain why the complex ontology for government budget reporting, XBRL, is actively being considered by both Congress and federal agencies. As implemented by the U.S. Securities & Exchange Commission for tagging corporate financial filings, XBRL has more than 18,000 standardized metadata tags based on Generally Accepted Accounting Principles (GAAP).

Although the focus here is on information sharing for democratic accountability, information sharing has also proved a powerful way to reduce costs across both governmental and non-governmental enterprises. Information sharing has often led to massive economies of scale and reduced costs in every part of the information cycle from creation to distribution to acquisition. Indeed, reducing costs is the major argument for standardization, not just standardized metadata, in almost every sphere where it is employed. The countless private and public organizations devoted to standardization are testimony to this. The irony is that high cost is the most popular excuse among politicians not to implement open government proposals. But it’s actually the weakest excuse and should be used as an argument for rather than against information sharing for democratic accountability.

The Politics of a Who Ontology for Democratic Accountability

Unfortunately, a natural human instinct in all spheres of endeavor is to avoid accountability. Americans, including elected public officials and the powerful political players who influence them, all love accountability for others but not themselves. But this information accountability NIMBYism is bad for our democracy. It is also probably the best explanation for why our ability to track the official actions of the politically powerful, including our elected representatives, remains so primitive.[3]

The reason tracking potential terrorists has been so effectively implemented across tens of thousands of government agencies from the local to the Federal level is not that tracking terrorists is technologically simpler and less expensive than tracking the politically powerful. Indeed, tracking the official actions of elected officials is technologically trivial and economically a pittance by comparison. Nor is it that potential terrorists are less resistant to being tracked than the politically powerful. It is that potential terrorists aren’t a powerful political group and cannot effectively engage in information accountability NIMBY politics.

The president needs to bravely set an example for all Americans by countering the natural tendency toward information accountability NIMBYism. No one else is better positioned, by setting a personal example, to usher in the new era of semantically enriched open government data. Since the president’s official and public actions are already so closely tracked by a large and sophisticated press corps, he also has the least to lose.

Given the conflict of interest politicians have in designing democratic information systems, government-wide information sharing for democratic accountability that moves beyond the president and his direct sphere of control involves many difficult governance issues, which are not addressed here. This is the realm where the NIEM governance model breaks down. More generally, it is the fundamental problem associated with open government public policies that are supposed to generate democratic accountability for the politically powerful as well as the weak. Ultimately, the governance problem can only be solved through the creation of new checks & balances institutions. By showing what’s technically and economically trivial to do to improve our democratic information systems, the president can cast a light on the centrality of governance issues.

Conclusion

The ability to easily track the official actions of elected representatives and the powerful political players who seek to influence them is essential to a healthy democracy. But endemic to democracy is that such powerful political players do not want their actions tracked. Since information is power, those with such power have no rational incentive to give it up. But for the sake of our democracy, they must be forced to do so. Progress, made possible by the emergence of semantic web technologies, should not be held hostage to their anti-democratic interests. The president, among all government officials, is best positioned to usher in the new era of technology-based democratic accountability by setting a personal example and pointing to the difficult governance issues that need to be addressed.

[1] The current versions of UCORE and NIEM, UCORE 2.x and NIEM 2.x, work together but are not fully integrated. NIEM 3.x and UCORE 3.x are expected to be fully integrated. NIEM/UCORE 3.x may also add RDFa and OWL functionality. Some UCORE 2.x implementations on military intranets have already added RDFa and OWL functionality.

[2] For reasons of privacy, unique personal identifiers for democratic accountability purposes should be different from identifiers, such as social security numbers and tax identification numbers, used for other purposes. The GUID may be well suited to take on this democratic function for not only major recipients of government largesse (e.g., government employees and contractors) but also major political players (e.g., registered lobbyists and large campaign donors).

[3] For a discussion of such NIMBYism in Congress, see J.H. Snider, The Dismal Politics of Legislative Transparency, Journal of Information Technology & Democracy, Spring 2009.

This paper draws heavily on J.H. Snider’s paper, Connecting the Dots for Democratic Accountability (Washington, DC: iSolon.org, October 22, 2010). Whereas this paper focuses on government-wide individual identifiers, that paper focused on government-wide organizational identifiers.

Downloads

Download the Paper

Authors

J.H. Snider

Image Source: Â© STR New / Reuters

The Dictators’ Digital Dilemma: When Do States Disconnect Their Digital Networks?

Sheetal D. Agarwal, Philip N. Howard and Muzammil M. Hussain — Wed, 05 Oct 2011 15:04:00 -0400

Executive Summary

When do governments decide to interfere with the Internet, and why? While many observers celebrate the creative use of digital media by activists and civil society leaders, there are a significant number of incidents involving government-led Internet shutdowns. Governments have offered a range of reasons for interfering with digital networks, employed many tactics, and experienced both costs and benefits in doing so.

When and why do states disconnect their digital networks is a principle question we examine in this paper. To answer this question, we built an event history database of incidents in which a regime went beyond mere surveillance of particular websites or users, and actually disconnected Internet exchange points or blocked significant amounts of certain kinds of traffic. All in all, there were 606 unique incidents involving 99 countries since 1995: 39 percent of the incidents occurred in democracies, 6 percent occurred in emerging democracies, 52 percent occurred in authoritarian regimes, and 3 percent occurred in fragile states. Then we developed three standardized typologies for the kinds of incidents being reported. First, we developed a category that iteratively helped define the case, and a typology of actions that states take against social media. Second, we developed a category for why they took that action, sometimes relying on third-party reports if the state simply denied any interference. Finally, we developed a category for the impact of the interference.

We find that overall more democracies participate in network interventions than authoritarian regimes. However, authoritarian regimes conduct shutdowns with greater frequency. After 2002, authoritarian governments clearly began using such interference as tool of governance. In recent years, even fragile states have interfered with domestic information infrastructure, usually as a last effort at maintaining social control.

Downloads

Download the Paper

Authors

Sheetal D. Agarwal
Philip N. Howard
Muzammil M. Hussain

Image Source: Â© Steve Crisp / Reuters

The Future of Spectrum

Jeffrey Rosen — Wed, 03 Aug 2011 10:11:00 -0400

Executive Summary

In recent years, growth in demand for wireless services has sparked a boom in the mobile phone and wireless data sector.[i] During the past four years, the number of mobile phone subscribers tripled,[ii] and the number of jobs in the telecommunications field has nearly quintupled.[iii] New, better, and faster mobile devices, such as tablets and smartphones, have created multi-billion dollar industries of their own, such as Google Android and the Apple iOS “app stores.”[iv] And those technologies have contributed to the dawning of an always-on, always-connected culture.

But this growing demand for mobile Internet access requires a growing amount of wireless radio spectrum, portending serious problems for the future. At the moment, the United States has designated 547 MHz of spectrum to wireless broadband services, but the Federal Communications Commission (FCC) predicts a need for 637 MHz of spectrum by 2013, and 822 MHz of spectrum by 2014.[v] Without more spectrum allocated to wireless Internet connectivity, America risks short-circuiting the mobile broadband revolution.

The National Broadband Plan proposes a solution. It sets forth a detailed plan to make 300 MHz of spectrum available for wireless broadband use within the next five years, and another 200 MHz in the five years after that.[vi] It seeks to achieve this freeing of spectrum by auctioning unused spectrum, lifting burdensome regulations to enable wireless broadband service in certain spectrum ranges, and reallocating spectrum from other services – notably broadcast television – to enable such spectrum to be used for wireless broadband.[vii] Though many of these provisions are controversial, the FCC has already done serious work to achieve these goals. If the FCC can achieve its goals to enable the growth of wireless broadband, America will be able to unlock the full potential of the wireless broadband revolution and realize the potential of a new wave of American innovation.

[i] Federal Communications Commission, Connecting America: The National Broadband Plan 78 (2010) [hereinafter National Broadband Plan].

[ii] Id.

[iii] Lawrence H. Summers, Remarks on the President's Spectrum Initiative As Prepared for Delivery (2010 June 28).

[iv] Robin Wauters, Report: Mobile App Market Will Be Worth $25 Billion By 2015 – Apple’s Share: 20 percent, TechCrunch.com, 2011 January 18, http://techcrunch.com/2011/01/18/report-mobile-app-market-will-be-worth-25-billion-by-2015-apples-share-20/.

[v] Federal Communications Commission, Mobile Broadband: The Benefits of Additional Spectrum 18 (2011) , available at http://download.broadband.gov/plan/fcc-staff-technical-paper-mobile-broadband-benefits-of-additional-spectrum.pdf (hereinafter Benefits of Additional Spectrum). [hereinafter Benefits of Additional Spectrum].

[vi] See National Broadband Plan, supra note 1, at 84.

[vii] Id.

Downloads

Download the Paper

Authors

Jeffrey Rosen

Image Source: Â© Luke MacGregor / Reuters

Making Public Community Media Accessible

J.H. Snider — Tue, 05 Jul 2011 15:18:00 -0400

Executive Summary

Early advocates for public, educational, and governmental access media (PEG), categorized here as “public community media,” envisioned that it would empower local civic groups through improved meeting coverage.

Unfortunately, the current reality has come up far short of the original promise. Economically, the technology proved prohibitively costly for civic groups to incorporate into their meetings. Politically, public officials had minimal incentive to make media available to civic groups over whom they didn’t have editorial control. The economic and political reasons reinforced each other in that the high cost of making public media available to civic groups, combined with the need for extensive government staff support, gave public officials a compelling excuse and means to exert the editorial control they sought.

Fortunately, new information technology is revolutionizing the economics and politics of meeting media, making it possible to cover face-to-face meetings at negligible cost and without the need for government controlled technical experts. Moreover, meeting participants can be empowered to participate in ways undreamed of by the early public community media advocates.

We can now imagine a world where webcasting a public meeting in a public building is as easy and inexpensive as flipping on a light switch, and where searching the contents of public meetings is as easy and inexpensive as a Google search.

We are clearly entering a new era of blended meetings, where meetings incorporate both face-to-face and cyberspace components in ways previously impractical or even impossible. However, for the democratic potential of this revolution in meeting technology to be realized, the public policy framework for public community media needs to be dramatically reformed.

So how should public media be redesigned to empower local civic groups? I lay out twelve recommendations:

Meeting Media Automation. Community media equipment should be designed to operate on a completely automated basis to eliminate the need for expertise and expensive labor.
Open and Extensible Meeting Software. The software to control community media equipment should be open and extensible so that additional meeting functionality can easily be added by third parties.
Interactive Meeting Media. Community meeting media should include not only broadcast TV but also interactive media, including voting devices, Wi-Fi, and flat screen TVs connected to the Internet, with remote participants able to largely replicate the public meeting media experience via the Internet.
Meeting-Friendly Public Spaces. Library meeting rooms, school auditoriums, and public access facilities should be designed to be meeting-friendly to civic groups.
Building Codes for Public Meeting Rooms. For government facilities substantially or primarily used for public meetings (“public meeting anchor institutions”), meeting technology should be included in building codes just like requirements for smoke detectors, exit signs, and electrical outlets.
Equity in Access to Meetings. To level the playing field between those with and without convenient physical access to public meetings, high quality remote access should be available.
Equity in Access to Meeting Equipment. Civic groups should be able to borrow inexpensive mics and clickers using the same infrastructure that allows citizens to borrow books from libraries.
Web-Centric Meeting Media. Community meeting media should be web, not cable TV, centric.
Civic Group Control of Meeting Media. Any government monopoly power on the use of public meeting media should be reduced as much as is feasible.
Checks & Balances Institutions for Meeting Places & Media. Public community meeting places and media should be implemented by checks & balances institutions.
Public Bodies Covered by State Open Meeting Laws. State open meeting laws should be revised to include in the definition of public bodies quasi-government bodies such as state mandated advisory groups, commissions, and ad hoc committees.
Federal Government Public Meetings. Although this essay focuses on local public media, the same principles apply to federal public media.

Downloads

Download the Paper

Authors

J.H. Snider

Image Source: Â© Hyungwon Kang / Reuters

Patents: A Singular Law for the Diversity of Innovation

Brian Kahin — Wed, 01 Jun 2011 10:03:00 -0400

Executive Summary

Not long ago, patents were viewed as a subject consigned to specialists with law and a technical background. Now branded as intellectual property, they have moved front and center in an economy dependent on innovation for sustainable growth. Yet as patents have expanded in number, power, and subject matter, they have drawn controversy as well as attention.

Today a unitary, deeply institutionalized legal system confronts an increasingly diverse technological and business environment. High demand, low standards of quality, and the extreme complexity of information have led not only to immense portfolios but to new varieties of strategic behavior. Large investments in product development are put at risk by the leveraging power of individual patents, and networked services and industry standards make especially attract targets because of the embeddedness and breadth of investment. While China has been criticized relentlessly for not respecting intellectual property rights, China now appears to be emulating the high-volume, low-quality U.S. model, but on its own terms.

Innovation is always a risky undertaking. In principle, patents help reduce risk by allowing an inventor a limited period to exclude others from exploiting the invention. However, the patent system can also create risk and uncertainty. Juries are notoriously unpredictable in patent cases, and judges have had difficulty applying the law, including the language of patent claims, with consistency. A lawyer’s opinion on whether a patent is valid may cost $15,000 or more – and another $15,000 or more for an opinion on whether a particular technology infringes the patent – but with no guarantees. But the greatest uncertainty facing innovators comes from the sheer volume of patents, most of which belong to somebody else.

The creation of the Court of Appeals for the Federal Circuit in 1982 was designed to make the patent system more consistent and predictable, at least at the appellate level, where there appeared to be inconsistency among the regional circuits. The Federal Circuit quickly became a champion of its specialty, making patents more powerful, easier to get, harder to attack, and available for a nearly unlimited range of subject matter. Low standards combined with the increased power of patent holders compounded the problem of indeterminacy. Patents grew in number and could be asserted aggressively after others had made huge investments in developing products that happened to infringe.

Controversies over the effects of this “strengthening” of the patent system simmered out of public sight until the Federal Trade Commission (FTC) and the Department of Justice held 24 days of hearings in 2002. In 2003, the FTC issued a landmark report that showed widespread concerns about patent quality and the functioning of the system, especially in information technology. The FTC report and a subsequent report by the National Academies spurred proposals for legislative reform.

Debate on patent reform often divides along the value chain (upstream vs. downstream) and across industries characteristics (complex vs. discrete technologies). Upstream research and design firms want leverage over downstream manufacturers, distributors, and retailers. Companies that assemble and market complex products are especially vulnerable to patents on the many thousands of individual components and functions that make up complex products.

The contrast between discrete and complex technologies is at the root of the so-called “patent paradox.” There are far fewer patents in pharmaceuticals than in information technology, but a single patent can enable huge investments to be made in clinical testing and regulatory approval – and can effectively secure the market value of a billion-dollar drug. Information technology is much more patent-intensive, but because there can be thousands of patentable functions in a very inexpensive product, the value of the technology protected by each individual patent is diluted. Yet individual patents can be valuable if inadvertently infringed, and IT firms have built up massive portfolios of patents for defensive purposes – to counter infringement claims. But “trolls” – or more politely, “patent assertion entities” – do not create products for the market. Undeterred by defensive portfolios or concerns about reputation, they can assert patents with impunity.^[1]

[1] See data collected by Patent Freedom: https://www.patentfreedom.com/research.html .

Downloads

Download the Full Paper

Authors

Brian Kahin

Ensuring Hardware Cybersecurity

John Villasenor — Wed, 04 May 2011 10:22:00 -0400

Executive Summary

More than ever, the electronic devices that are critical to everyday life, to the larger infrastructure, and to national defense are dependent on increasingly sophisticated semiconductor integrated circuits, also referred to as “chips”. For example, laptop computers and tablets, smartphones, the financial system, the Internet, aircraft flight controls, automobile antilock braking, the power grid, and an almost endless list of other devices and systems can be trusted to run properly only if the chips they contain are free of hidden malicious circuits inserted during the design or manufacturing process.

The combination of continued chip technology advances and an unprecedented level of globalization in the semiconductor industry has spurred enormous changes in the way chips are designed, manufactured, and used. These changes bring many benefits to the consumer including lower prices and faster time to market for products and services, but they have also created a widening set of opportunities for would-be attackers to insert malicious circuits during the chip design process that could be used to launch a hardware attack.

Despite the potentially devastating impact that a large-scale hardware attack could have on commerce, defense, and government function, the need to proactively address hardware security remains widely underappreciated. This document explains the causes and nature of the hardware security threat and outlines a multipronged approach to address it involving 1) a change in design practices within the semiconductor industry, 2) the establishment of a national-level capability to coordinate a quick response to an attack, 3) improved testing procedures to detect corrupted chips before they are placed into products, and 4) the inclusion of built-in defenses into chips to identify and thwart attacks as they occur.

Introduction

Modern electronic semiconductor integrated circuits are marvels of miniaturization, capable of packing into a circuit approximately the size of a nickel an amount of computational power far exceeding that delivered by an entire roomful of computers in the 1960s. By enabling the devices and networks that allow information to be almost instantly acquired, processed, searched, and shared, integrated circuits, or “chips”, have revolutionized the technology landscape. In doing so, they have played a key behind-the-scenes role in reshaping social interactions and the political landscape as well.

The semiconductor industry has become an enormous global enterprise that generated nearly $300 billion in sales in 2010. There are approximately 1500 companies in the world today engaged in chip design. The full chip ecosystem, which includes chip designers and manufacturers, companies that use the resulting chips in products, and individuals, corporations, and governments that in turn purchase these products, relies on the assumption that the chips at the core of these products have integrity.

Until the very recent past, this has been a reasonable assumption. However, continued changes in the dynamics of the global semiconductor market have made it not only possible but inevitable that chips that have been intentionally and maliciously altered to contain hidden “Trojan” circuitry will be inserted into the supply chain. These Trojan circuits can then be triggered months or years later to launch attacks. It is imperative to put into place systems that can minimize the likelihood, number, and consequences of these attacks.

What is a Hardware Attack?

With respect to chips and the devices in which they reside, tasks such as performing an Internet search, editing a document on a computer, making a mobile phone call, and conducting a financial transaction rely on a complex interplay between software and hardware. “Software” refers to the set of instructions that describe how a task is performed, while “hardware” refers to the circuits that actually do the work to perform the task.

While software can be replaced, updated, altered, and downloaded from the Internet, chip hardware generally can’t be changed after the chip leaves the factory. Thus, while malicious software can be created and disseminated by anyone with a computer and access to the Internet, malicious hardware can only be inserted by someone who can access and alter the design for a chip before it is manufactured and placed in a product.

Once malicious hardware has been built into a chip, a hardware attack can be initiated and act in a wide variety of ways. An attack can be internally triggered, based, for example on the arrival of a particular calendar day. Alternatively, an external trigger could be hidden within data sent by an attacker. More complex hybrid triggers could also be used. For example, a malicious circuit hidden within a GPS chip could be configured to attack only when the chip is located in a specific geographical area after a certain date.

There are multiple forms of potential attacks. In an overt attack, the malicious hardware could cause the device containing the corrupted chip to either cease functioning altogether or to continue to operate but in an obviously impaired manner. The existence of a problem would be clear, though its cause would not. In a personal electronics device such as a mobile phone such an attack could be nothing more than an inconvenience. If conducted on a large scale on thousands of chips within a critical portion of the national infrastructure, this form of attack could be devastating.

In contrast with an overt attack, in a covert attack the appearance of normal operation is maintained while malicious actions are quietly being performed in the background. For example, a corrupted chip within a communications system could be caused to send copies of confidential data to a third-party destination.

A third form of attack leaves the device operating, but introduces corruptions into the data. As an example, consider an attack hidden in a GPS chip with a location-based trigger that left the GPS functioning accurately until it was located in a certain geographical region. Upon receiving this geographical trigger, it could act by shifting GPS locations by a few hundred feet. An attack of this form would be extremely difficult to detect in advance and could have significant consequences across a wide range of application scenarios. There can also be attacks that exploit a hybrid of malicious hardware and software. Malicious hardware hidden within a chip could be triggered months or years later to open a back door allowing the installation of malicious software which in turn could launch the attack.

The Growth of Chip Complexity – and Vulnerability

It is the very size and complexity of modern chips that creates the vulnerabilities that make the insertion of malicious hardware possible. Gordon Moore, a co-founder of Intel, famously predicted in a 1965 paper that the amount of functionality that can be built into a single chip would double approximately every two years. Remarkably, “Moore’s Law” has stood over a time span now approaching half a century. This level of sustained growth is made possible in large part due to continued advances in the ability to miniaturize the size of the structures within a chip. In the 1970s the smallest structure sizes that could be created were several thousand nanometers in width (one nanometer is one one-billionth of a meter). Today, the most advanced semiconductor manufacturing methods are able to create structures as small as a few tens of nanometers – hundreds of times smaller than what was possible in the 1970s.

The existence of chip vulnerabilities today can be traced in part to the history of chip design, and in this respect has analogies with the growth of the Internet. As is well known, the Internet was originally designed by a small community of researchers as an open environment accessed by a small number of trusted users. This assumption of trust became problematic with enormous growth in the size of the Internet, and with it, an increasing number of users willing to exploit weaknesses in the Internet for malicious purposes.

Chip design too has its heritage in small, self-contained teams with a shared interest in the success of the chip and in which there was no reason to question trust. Many of the practices used in chip design and many of the protocols for processing data and for communicating data among parts of a chip were established in that environment. In today’s world, however, the design process for a single chip can involve contributions hundreds of people, many of whom may be employed by third party companies that simply provide functional blocks and who have little or no stake or interest in the success of the chip.

The Globalization of Chip Design

The process of creating a chip includes the two sequential phases of design and manufacturing. Design involves determining the functionality of the chip and mapping that functionality into a description in terms of electronic circuits. Manufacturing refers to physically producing the chip containing these circuits. Insertion of malicious hardware during manufacturing is very difficult because of the likelihood that the insertion process itself will lead to impairments that would be detected during post-manufacturing testing. For an attacker, the low hanging fruit lies in the design process, where there is the potential to create malicious circuits and bury them within the much larger set of healthy circuits in a non-disruptive manner.

The number of people who are in a position to access and therefore potentially compromise chip designs is vastly smaller than the number of people who could create malicious software. But, in absolute terms, thanks in large part to outsourcing, it is still very large, with many hundreds of thousands of people around the world directly employed in the chip design industry.

Chip design today relies heavily on outsourcing. Although a complex chip is a single, physically small device, it contains many different functional areas, called “blocks,” that perform different tasks. A chip used in a smartphone, for example, may have a set of functional blocks devoted to receiving a wireless signal, processing that signal to extract the data it contains, decoding that data to produce audio and video signals, and sending those signals to a speaker and display screen. Much as an architectural firm charged with designing an enormous shopping complex might subcontract out portions of the design, a company overseeing the design of a complex chip typically designs some portions in house but obtains designs for other portions from third parties. While outsourced chip manufacturing has been common for several decades, the use of outsourcing in chip design has accelerated dramatically in the last half decade. As with all outsourcing, the primary driver for this is economic. Labor costs are much lower in places like India and China, and companies located in high-cost labor markets cannot remain competitive if they rely exclusively on in-house chip designers. The combination of growth in both complexity and outsourcing means that the number of people with access to the design for a single chip during its development can easily number in the hundreds.

Given this landscape, there are multiple potential vectors for the insertion of malicious hardware. One possibility is that a company performing outsourced design services could intentionally provide a corrupted design with the full knowledge and participation of the company management. Alternatively, the design services company could act in good faith, but could store the designs on weakly secured networks, enabling the designs to be accessed and altered by an outside party. It is also possible for one or more individuals within the design services company to corrupt a design without the knowledge of their colleagues or managers.

Finally, malicious hardware insertion can also occur in non-outsourced portions of the design. An “inside job” perpetrated by one or more rogue employees employed at the home company overseeing the entire chip design could be particularly hard to detect because of the higher level of access and knowledge regarding the overall chip that these employees often possess. As an alternative, network vulnerabilities could be exploited to access and corrupt non-outsourced designs.

The Challenge of Testing

In an ideal world, corrupted designs would be detected, regardless of their source. However, the sheer complexity of modern chips greatly impedes such detection. While extensive – but not exhaustive - testing is performed during the design and manufacturing process, the goal of this testing is to confirm that a chip is behaving as expected. The testing procedures are very good at identifying accidental design flaws, but are poorly suited to ferreting out intentionally hidden malicious circuitry.

Consider the following example: Suppose that a company outsources the design for a block of the chip that is supposed to add the number six to any input. During testing, if 20 is provided to this block, the block outputs 26. When 127 is provided, the block outputs 133. One hundred thousand more inputs are provided, and in every case, the result comes back correct. This block will be deemed to have passed functional testing. But the block could have a hidden circuit triggered by an input with value 126,321,204. When that input – and that input alone – arrives, an attack is launched. Because testing can’t possibly be exhaustive, this input will never be encountered until it is provided months later by an attacker.

Towards a Comprehensive Solution

If a significant hardware attack were to be launched today, we would be ill-equipped to respond. In stark contrast with the large amount of attention and resources being directed to ensuring software security, efforts to address the potential impact of the contamination of the commercial chip supply by malicious hardware are in their infancy at best. The following steps could go a long way toward reducing the likelihood and impact of hardware attacks:

A change in design practices within the semiconductor industry

Current design practices in the semiconductor industry should be modified to specifically recognize and address the potential for malicious hardware insertion. First, companies engaged in chip design should adopt a need-to-know partitioning of information, much as occurs within the Department of Defense or in companies engaged in defense contract work. A designer working on a portion of a chip devoted to receiving wireless data does not need access to the internal details of a portion of the chip that processes video for display on the screen. But, in many companies, the barriers that separate design access are either nonexistent or insufficiently high. Second, companies engaged in chip design should recognize the existence of a real and significant threat from malicious hardware – a threat that is not generally even on the radar screen in any meaningful way at most such companies. This could lead to a more careful scrutiny of third-party suppliers as well as to measures that would reduce the odds that designs could be compromised under their own roofs.

Establishment of a national-level capability to coordinate a quick response to an attack

Currently, we do not have any national level capability to respond to an attack. This would greatly impede our ability to respond in an agile manner. For example, if an attack significant enough to require a national-level response were to occur today, it is not clear which governmental entity would oversee the response, and valuable time would be lost in creating the appropriate organizational framework and in establishing the appropriate lines of communication and information flow.

To avoid this scenario, the entity that would be charged with overseeing the response to a hardware attack should be identified or created preemptively. Procedures should be put in place for reporting an attack and for engagement with the appropriate companies and governmental organizations.

In the event of an attack, it would also be critically important to be able to rapidly identify the other chips containing designs received from a known supplier of corrupted hardware. Currently, this identification could take many weeks or longer – precious time that could confer significant advantages to an attacker. However, a government-managed database of suppliers could allow this identification to be nearly instantaneous. It is not feasible to simply ask companies to furnish the government a list of suppliers every time they unveil a new chip, as this information is typically considered highly proprietary. However, with proper design using multiple layers of asymmetric encryption, companies could provide supplier information to the government that would enable rapid tracing while also avoiding the unnecessary disclosure of proprietary information.

Improved testing procedures to detect corrupted chips before they are placed into products

Today’s commercial chip testing procedures are designed to identify accidental design flaws, not to discover intentionally hidden attacks. By developing new testing procedures that are specifically designed to look for attacks, the odds that corrupted hardware could escape pre-deployment testing would be significantly reduced. Fortunately, the research arm of the Department of Defense, the Defense Advanced Research Projects Agency (DARPA), has been funding a program for the past several years directed to testing untrusted circuits to look for malicious hardware. The solutions developed under this program have the potential to play a significant role in reducing the odds that compromised hardware will be deployed.

Inclusion of built-in defenses into chips to identify and thwart attacks as they occur.

While pre-deployment testing is extremely important, it cannot be relied on to find all instances of malicious hardware. Indeed, as stated in a 2005 report by the Defense Science Board Task Force on High Performance Microchip Supply, “[E]lectrical testing and reverse engineering cannot be relied upon to detect undesired alterations in military integrated circuits” (Defense Science Board Task Force on High Performance Microchip Supply, 2005). This challenge was also noted by U.S. Deputy Secretary of Defense William J. Lynn III in a September/October 2010 article in Foreign Affairs, who noted with respect to hardware that “[t]ampering is almost impossible to detect.” (Lynn, September 2010)

Given the inevitability that some number of compromised chips will slip past the testing process, it is important to build defenses into chips that can identify and respond to attacks within milliseconds (Villasenor, “The Hacker in Your Hardware”, Scientific American, 2010). This could be accomplished by adding a modest amount of circuitry specifically charged with the task of monitoring the behavior of the chip and identifying behavior that may be indicative of an attack. When an attack is discovered, the offending portion of the chip could be identified and quarantined, and a notification sent to other devices containing similar circuits. Clearly, this requires that the circuits doing the policing are themselves free of corruption – which is more likely to occur if the “policing” portions of the chip are designed in-house using a very small group of highly trusted engineers.

In sum, while it is impossible to completely eliminate the potential for hardware attacks, the above measures, in combination, could significantly reduce the odds of a successful attack.

Downloads

Download the Paper

Authors

John Villasenor

Universal Service Fund Reform: Expanding Broadband Internet Access in the United States

Jeffrey Rosen — Tue, 05 Apr 2011 10:51:00 -0400

Executive Summary

Two-thirds of Americans have broadband Internet access in their homes.[1] But because of poor infrastructure or high prices, the remaining third of Americans do not. In some areas, broadband Internet is plainly unavailable because of inadequate infrastructure: More than 14 million Americans – approximately 5 percent of the total population – live in areas where terrestrial (as opposed to mobile) fixed broadband connectivity is unavailable.[2] The effects of insufficient infrastructure development have contributed to racial and cultural disparities in broadband access; for example, terrestrial broadband is available to only 10 percent of residents on tribal lands.[3]

Even where terrestrial broadband connectivity is available, however, the high price of broadband service can be prohibitive, especially to lower income Americans. While 93 percent of adults earning more than $75,000 per year are wired for broadband at home, the terrestrial broadband adoption rate is only 40 percent among adults earning less than $20,000 annually.[4] These costs also contribute to racial disparities; almost 70 percent of whites have adopted terrestrial broadband at home, but only 59 percent of blacks and 49 percent of Hispanics have done the same.[5]

America's wireless infrastructure is better developed, but many Americans still lack wireless broadband coverage. According to a recent study, 3G wireless networks cover a good portion of the country, including 98 percent of the United States population,[6] but certain states have dramatically lower coverage rates than others. For example, only 71 percent of West Virginia's population is covered by a 3G network.[7] Wireless providers will likely use existing 3G infrastructure to enable the impending transition to 4G networks.[8] Unless wireless infrastructure expands quickly, those Americans that remain unconnected may be left behind.

Though America is responsible for the invention and development of Internet technology, the United States has fallen behind competing nations on a variety of important indicators, including broadband adoption rate and price. According to the Organization for Economic Cooperation and Development's survey of 31 developed nations, the United States is ranked fourteenth in broadband penetration rate (i.e. the number of subscribers per 100 inhabitants); only 27.1 percent of Americans have adopted wired broadband subscriptions, compared to 37.8 percent of residents of the Netherlands.[9]

America also trails in ensuring the affordability of broadband service. The average price for a medium-speed (2.5Mbps-10Mbps) Internet plan in America is the seventeenth lowest among its competitor nations. For a medium-speed plan, the average American must pay $38 per month, while an average subscriber in Japan (ranked first) pays only $22 for a connection of the same quality.[10]

The National Broadband Plan (NBP), drafted by the Federal Communication Commission and released in 2010, seeks to provide all Americans with affordable broadband Internet access.[11] Doing so will not be cheap; analysts project that developing the infrastructure necessary for full broadband penetration will require $24 billion in subsidies and spending.[12] President Obama’s stimulus package has already set aside $4.9 billion to develop broadband infrastructure,[13] and some small ongoing federal programs receive an annual appropriation to promote broadband penetration.[14] However, these funding streams will only account for one-third of the $24 billion necessary to achieve the FCC's goal of full broadband penetration.[15] Moreover, developing infrastructure alone is not enough; many low-income Americans are unable to afford Internet access, even if it is offered in their locality.

To close this funding gap and to make broadband more accessible, the National Broadband Plan proposes to transform the Universal Service Fund – a subsidy program that spends $8.7 billion every year to develop infrastructure and improve affordability for telephone service – into a program that would do the same for broadband Internet.

^{^[1]} Federal Communications Commission, Connecting America: The National Broadband Plan 23 (2010) [hereinafter National Broadband Plan].
^{^[2]} Id. at 10.
^{^[3]} Id. at 23.
^{^[4]} Id.
^{^[5]} Id.
^{^[6]} Id. at 146.
^{^[7]} Id.
^{^[8]} Id.
^{^[9]} Organization for Economic Cooperation and Development, OECD Broadband Portal, OECD.org, (table 1d(1)) (last accessed Jan. 28, 2011).
^{^[10]} Id. (table 4m) (last accessed Jan. 28, 2011).
^{^[11]} National Broadband Plan, supra note 1, at 9-10.
^{^[12]} Id. at 136.
^{^[13]} Id. at 139.
^{^[14]} Id.
^{^[15]} Id.

Downloads

Download the Full Paper

Authors

Jeffrey Rosen

Image Source: Donald E. Carroll

The Terms They Are A-Changin'...: Watching Cloud Computing Contracts Take Shape

Simon Bradshaw, Christopher Millard and Ian Walden — Tue, 01 Mar 2011 13:45:00 -0500

EXECUTIVE SUMMARY

Many web services are examples of cloud computing, from storage and backup sites such as Flickr and Dropbox to online business productivity services such as Google Docs and Salesforce.com. Cloud computing offers a potentially attractive solution to customers keen to acquire computing infrastructure without large up-front investment, particularly in cases where their demand may be variable and unpredictable, as a means of achieving financial savings, productivity improvements and the wider flexibility that accompanies Internet-hosting of data and applications.

The greater flexibility of a cloud computing service as compared with a traditional outsourcing contract may be offset by reduced certainty for the customer in terms of the location of data placed into the cloud and the legal foundations of any contract with the provider. There may be unforeseen costs and risks hidden in the terms and conditions of such services.

This document reports on a detailed survey and analysis of the terms and conditions offered by cloud computing providers.

The survey formed part of the Cloud Legal Project at the Centre for Commercial Law Studies (CCLS), within the School of Law at Queen Mary, University of London, UK. Funded by a donation from Microsoft, but academically independent, the project is examining a wide range of legal and regulatory issues arising from cloud computing. The project's survey of 31 cloud computing contracts from 27 different providers, based on their standard terms of service as offered to customers in the E.U. and U.K., found that many include clauses that could have a significant impact, often negative, on the rights and interests of customers. The ease and convenience with which cloud computing arrangements can be set up may lull customers into overlooking the significant issues that can arise when key data and processes are entrusted to cloud service providers. The main lesson to be drawn from the Cloud Legal Project’s survey is that customers should review the terms and conditions of a cloud service carefully before signing up to it.

The survey found that some contracts, for instance, have clauses disclaiming responsibility for keeping the user’s data secure or intact. Others reserve the right to terminate accounts for apparent lack of use (potentially important if they are used for occasional backup or disaster recovery purposes), for violation of the provider’s Acceptable Use Policy, or indeed for any or no reason at all. Furthermore, whilst some providers promise only to hand over customer data if served with a court order, others state that they will do so on much wider grounds, including it simply being in their own business interests to disclose the data. Cloud providers also often exclude liability for loss of data, or strictly limit the damages that can be claimed against them – damages that might otherwise be substantial if a failure brought down an e-commerce web site.

Although in some U.S. states, in E.U. countries and in various other jurisdictions the validity of such terms may be challenged under consumer protection laws, users of cloud services may face practical obstacles to bringing a claim for data loss or privacy breach against a provider that seems local online but is, in fact, based in another continent. Indeed, service providers usually claim that their contracts are subject to the laws of the place where they have their main place of business. In many cases this is a US state, with a stipulation that any dispute must be heard in the provider’s local courts, regardless of the customer’s location.

Perhaps the most disconcerting discovery of the Cloud Legal Project’s survey was that many providers claimed to be able to amend their contracts unilaterally, simply by posting an updated version on the web. In effect, customers are put on notice to download lengthy and complex contracts, on a regular basis, and to compare them against their own copies of earlier versions to look for changes.

The cloud computing market is still developing rapidly, and potential cloud customers should be aware that there may be a mismatch between their expectations and the reality of cloud providers' service terms, and be alive to the possibility of unexpected changes to the terms.

Downloads

Download the Full Paper

Authors

Simon Bradshaw
Christopher Millard
Ian Walden

Image Source: Natalie Racioppa

Immigration and High-Impact, High-Tech Entrepreneurship

Zoltan J. Acs and David M. Hart — Mon, 07 Feb 2011 00:00:00 -0500

EXECUTIVE SUMMARY

Our study shows that the founding teams of about 16 percent of a nationally representative sample of high-impact, high-tech companies – the kind of company that is most critical for long-term economic growth – include at least one immigrant. These immigrant entrepreneurs are deeply-rooted in the U.S.; about 77 percent, for instance, are U.S. citizens. Most are well-educated and have substantial professional experience. Our evidence does not allow us to rule out the possibility that the immigrant entrepreneurs “crowd out” comparable natives, but we incline toward the view that immigrants and natives complement one another. We outline three policy options that might expand the pool of potential high-impact, high-tech immigrant entrepreneurs over the long-term: clearing the green card backlog, easing the pathway from student visa to work visa to green card, and creating a “point system” for a limited number of unsponsored green card applicants.

High-Impact, High-Tech Companies: One Key to Strong Economic Performance

Economic growth is an essential objective of the U.S. government, perhaps the essential objective. Invariably, when times are hard, voters turn incumbents out of office, as they did in 2010. Challengers know that keeping the campaign focus on the economy will pay dividends. The 1992 Clinton presidential campaign motto — “It’s the economy, stupid” — sums it up.

Governing is harder than campaigning. It may be “stupid” for a challenger to ignore the economy, but the incumbent who is not sure what to do to spur growth does not deserve that label. The American economy is a huge, complex, and rapidly evolving system that is increasingly open to global forces. Changing its trajectory is tricky, and experts disagree about how to do it.

In the short-term, the debate focuses on macroeconomic policy tools, such as interest rates and budget deficits, that may put under-utilized resources, including idled workers, back to work. In the long-term, however, the challenge is to expand the resources that are available and to enhance the creativity and efficiency with which they are used. We need microeconomic policies that stimulate innovation and productivity along with sound macroeconomic policies.

Even in hard times, which demand short-term responses, we should not neglect the long-term challenges. Our research contributes to this agenda by identifying a potential point of leverage for enhancing the economy’s long-term growth prospects. We draw on earlier work (Acs, Parsons, & Tracy 2008) that shows that a small share – just 3-4 percent — of all companies in the U.S. are “high-impact companies” that are responsible for most of the country’s economic growth and job creation.

We also draw on prior work that suggests that high-tech companies are particularly important for long-term growth (Acs et al. 2009). The products of these companies help other companies improve their productivity and give consumers new choices. Less than 10 percent of the high-impact companies operate in high-tech sectors. High-impact, high-tech companies therefore comprise less than 1 percent of all the companies in the U.S. If this share could be raised, even a little bit, the economy might benefit a lot.

Immigrants Found a Substantial Proportion of High-Impact, High-Tech Companies

Most of America’s high-impact, high-tech companies are relatively young, and the vast majority are still owned by their founders. Our project seeks to learn more about these entrepreneurs. Although many factors (such as the availability of start-up capital, access to promising markets, etc.) shape the success of each of these businesses, the founder’s skill set and insight certainly play a big part in it.

Our particular question about the founders is where they come from. Many observers of the U.S. high-tech industry believe that immigrant entrepreneurs play a special role in the industry. Some suggest that immigration policy ought to target potential high-tech entrepreneurs. Such a policy might, in principle, provide the kind of leverage on long-term economic growth that we are looking for.

Research has lagged behind the public dialogue on this subject. Although a number of studies (e.g. Saxenian 1999, Ballou et al. 2007, Wadhwa et al. 2007) have sought to pin down the share of immigrants among entrepreneurs, none focused on high-impact, high-tech companies on a national basis.

We carried out a telephone survey of a nationally representative sample of such companies. These companies, which were contacted in late 2008, had doubled in size between 2002 and 2006. We followed up the survey by conducting a set of interview-based case studies of companies founded by individual immigrant entrepreneurs, individual native-born entrepreneurs, and mixed teams of immigrant and native-born entrepreneurs.

Our survey discovered that the founding teams of about 16 percent of the high-impact, high-tech companies in our sample include at least one immigrant entrepreneur, defined as a founder of the company who was born outside of the U.S. This figure is on the low end of the range reported by prior research, but it is still a substantial proportion. Slightly more than half of the companies in our sample were founded by a single individual; the approximately 1300 companies have about 2000 founders. Of these 2000 founders, about 13 percent were born outside of the U.S.

Who Are the Immigrant Entrepreneurs?

The vast majority of the immigrant entrepreneurs in our sample are strongly rooted in the U.S. [i] The average duration of their stay in this country is more than twenty-five years. Only about 25 percent were reported to have been in the U.S. for less than fifteen years. About 77 percent are U.S. citizens. They came to the U.S. from all over the world. Fifty-five countries of origin are represented in the sample.[ii] India tops the list, accounting for about 16 percent of the founders. (See Table 1 - available in the downloadable paper version.)

The immigrant entrepreneurs are well-educated. Roughly 55 percent of them hold a masters degree or doctorate. Immigrant founders are more than twice as likely as native-born founders to hold a doctorate and are much more likely to hold a masters degree as well. One important reason for this difference is that many of the immigrant founders came to the U.S. for graduate education. Two-thirds of them received their highest degree in this country.

Like most successful entrepreneurs, the immigrant entrepreneurs in our sample had significant work experience. More than half of them had been in the U.S. at least ten years before founding their companies. In addition, many are serial entrepreneurs. Although our survey did not ask whether the founders had started a company before, the founders of more than half of the companies in our case studies had done so.

The 13 percent share of immigrant entrepreneurs in our survey sample is roughly the same as the foreign-born share of the entire U.S. population today. However, considering that the vast majority of the immigrant entrepreneurs in our sample have been in the U.S. for two or more decades and are highly educated, a more appropriate comparison is to the foreign-born share of the U.S. population holding a bachelor’s degree or in the science, technology, engineering and mathematics (STEM) workforce in 1990, both of which were about 10 percent. These baseline populations have grown since 1990, which means that the pool of potential high-impact immigrant entrepreneurs has grown as well.

The case studies tell stories that breathe life into these statistics. For instance, one of the immigrant entrepreneurs whom we interviewed is a prolific inventor. He came to the U.S. because he felt unable to pursue the business opportunities opened up by his inventions in his country of origin. He had founded several companies prior to the one in our sample and, at the time of our interview, he was operating what amounted to a personal incubator for several nascent companies based on more recent inventions.

Two other immigrant entrepreneurs in our case study companies are faculty members. One came to the U.S. to get his doctorate, the other as a post-doctoral fellow. Both had lengthy academic research careers before pursuing entrepreneurship, although in both instances, the company in our sample was not the first one that the entrepreneur had started. In both instances as well, the faculty entrepreneur teamed up with a student; one of these students hails from his professor’s country of origin, while the other student is native-born.

Do Immigrant Entrepreneurs “Crowd Out” Native-Born Entrepreneurs?

Although entrepreneurs like these are extraordinary individuals, we do not know for sure that the entrepreneurial opportunities that they recognized and their strategies for pursuing them were unique. After all, they came to the U.S. to take advantage of the institutional environment here, to serve in America’s great universities in the case of the academics and to access risk capital and risk-taking business partners in the case of the prolific inventor. Perhaps it is the institutional environment that is truly unique and not the entrepreneurs. To put it another way, if these individuals had not come to the U.S. and had not founded high-impact, high-tech companies, perhaps a native-born entrepreneur, drawing on the same institutional environment, would have done so instead.

One way that our research gets at this difficult-to-study issue is by looking at whether companies founded by one or more immigrants are engaged in different lines of business than those founded by natives. Neither the quantitative data from the survey nor the qualitative data from the case studies suggest that this is the case. Nor do we find that immigrant-founded companies are bigger, more likely to engage in research and development, or more likely to hold patents than the native-founded companies in the sample, once we control for other factors. This work suggests that we cannot rule out the “crowding out” hypothesis.

There is a complication in this analysis, however, which has to do with one of the factors that we are controlling for, the founders’ education. Companies with highly-educated founders are different in all of these respects. If we remove the control for education, immigrant-founded companies appear to be different as well, which is not surprising, since we know that immigrant founders tend to be more highly-educated than native-born founders.

Which of these variables is the right one to focus on? That is a matter of interpretation. Most of the immigrant entrepreneurs were educated in the U.S. If one believes that their academic success, which anticipates their later entrepreneurial success, was due to their innate talent, or to their training before they came here, then one should privilege immigration over education in the analysis. This interpretation assumes that equally gifted native-born students would not have been available to take the immigrant entrepreneurs’ places in school, and thus there was no “crowding out.” On the other hand, if it was their U.S. education that was critical to the immigrants’ success as entrepreneurs, native-born substitutes might have done equally well, whether or not these hypothetical substitutes were as gifted as the immigrants before they got their education.

One difference that holds up even when we control for education is that immigrant-founded companies are more likely to report that they have a strategic relationship with a company in another country than their native-founded counterparts. Our interviews indicate that these relationships sometimes involve inputs, such as product development services, and sometimes outputs, such as sales to foreign customers. The interviews also suggest that the foreign partner is usually in the immigrant entrepreneur’s country of origin. Presumably such a relationship would not have been created by our hypothetical native-born substitute.

How should we interpret this finding? On the one hand, it may be the case that the strategic relationships that immigrant-founded companies maintain with their foreign partners were essential to their success and possibly even their survival. If this is so, any alternative strategy that might have been pursued by our hypothetical native-born substitute would have been inferior. On the other hand, perhaps these relationships were matters of choice without a material impact on the company. In that case, the hypothesis that the immigrant entrepreneurs are “crowding out” native-born entrepreneurs, once again, cannot be ruled out. Unfortunately, our data do not provide a definitive answer. More research is needed.

We found one more interesting difference in our study that has to do with the composition of founding teams. (Slightly less than half of the companies in the sample were founded by teams, rather than individuals.) Immigrant founders were more likely than their white, native-born counterparts to team up with other “outsiders,” not just other immigrants, but also female and U.S. minority entrepreneurs. Perhaps the availability of immigrant entrepreneurs as teammates creates entrepreneurial opportunities for members of these under-represented groups that would not have been created in their absence.

Policy Options for Fostering High-Impact, High-Tech Immigrant Entrepreneurship

While our study permits a range of interpretations, we incline toward the view that immigrant entrepreneurs complement, rather than “crowd out,” native-born entrepreneurs. This position finds support in other studies (e.g. Hunt 2010, No & Walsh 2010, Ortega & Peri 2009) that focus on scientific publications and patenting as well as high-tech entrepreneurship. While only a small fraction of immigrants pursues entrepreneurship, much less succeeds at high-impact, high-tech entrepreneurship, the economic leverage of this activity is such that we believe it ought to be considered by immigration policy-makers.

The objective of this aspect of immigration policy should be to expand the pool of potential high-impact, high-tech entrepreneurs, rather than to seek to identify promising individuals. Entrepreneurship is by its nature a high-risk venture, fraught with failure, even for the most well-endowed aspirants. If our conjecture that immigrant entrepreneurs are complements to, rather than substitutes for, native-born entrepreneurs is correct, more tries by immigrant entrepreneurs will produce more successes. So, the policy goal should be to encourage as many tries as possible, not to “pick winners.”

Option 1: Clear the Employment-Based Green Card Backlog

The most important fact that policy-makers need to bear in mind is the long lag between immigration and impact. Both higher education and extensive work experience contribute to entrepreneurial success, and both take time to acquire. The entrepreneurs in our sample reflect the flow of immigrants two or more decades ago, rather than in the last few years.

One way to expand the pool of potential immigrant entrepreneurs in the short- to medium-term is to reduce the constraints on educated and experienced workers who are already in the process of immigrating to the U.S. Jasso et al. (2010) estimate that about a half-million applicants were waiting for employment-based green cards (EB-1, EB-2, and EB-3 categories) in 2006, an eight to ten year backlog. Until an applicant reaches the final stage of the process, she must remain an employee of her sponsor. Uncertainty about the length of the wait and the outcome of the process, along with the need for an employer sponsor, surely frustrate the entrepreneurial ambitions of some applicants.

Although some of the backlog is the result of processing delays, a more fundamental cause is the legal limit on the number of green cards that can be issued overall and to applicants from any given country each year. This policy ignores the size of a country’s population and the size and composition of its applicant pool. Applicants from large countries such as India and China must wait longer than those from smaller countries. For example, in the EB-2 category, for professionals with advanced degrees, applicants from these two countries had to have applied for green cards by mid-2006 in order to have their applications processed in January 2011. In the EB-3 category, professionals with bachelor’s degrees and skilled workers, Indian applicants from early 2002 were still being processed. (U.S. Department of State, 2011)

Clearing this backlog would require legislation to lift, at least temporarily, the quotas for employment-based applicants. To prevent the backlog’s re-emergence in the future, the flow of presumptive immigrants on temporary visas would need to be better matched to the availability of employment-based green cards. One possibility for accomplishing this goal would be to raise the global total of employment-based green cards, while reforming temporary visa programs. In parallel, policymakers might consider a system of awarding green cards that gives preference on the basis of criteria that are better correlated than country of origin with the potential contributions of prospective immigrants to high-impact entrepreneurship and other economically-valuable activities. In the meantime, the requirement for continued employment with the applicant’s sponsor during the green card process might be relaxed.

Option 2: Ease the Pathway from Student Visa to Work Visa to Green Card

A second policy option focuses on simplifying and clarifying the immigration pathway for potential high-impact entrepreneurs. The current policy sends mixed signals to many holders of student and work visas. In principle, most are obligated to leave the U.S. when their visas expire; in practice, many find ways to stay. They extend, shift, adjust, overstay, or simply wait for the rules to change. Uncertainty and confusion takes a toll on them along the way.[iii]

One approach to simplification would be to automatically allow foreign students (or some subset of them) to seek jobs in any field for a limited period of time after completing their degrees.[iv] Those who are successful in doing so and then remain employed might not only maintain their right to stay in the U.S., but also be put on a fast track for permanent admission and be exempted from the requirement to have an employer sponsor. A threshold for income or job quality might be applied to assure that these former students are continuing to progress along a promising career path before they receive their green cards.

This approach differs from the proposal to “staple the green card to the diploma” of foreign students in certain fields. It would provide more assurance that presumptive immigrants acquire work experience as well as education. It would also avoid inducing the enrollment of poor-quality foreign students in U.S. higher education institutions simply to obtain green cards. It differs as well from using temporary work visa programs, such as the H-1B, as a pathway to the green card. Such programs should serve their intended purpose, which is to alleviate temporary labor shortages in high-skill occupations.

Option 3: Create a “Point System” That Allows Potential Entrepreneurs to Get Green Cards

A third and much more challenging option would be to create a “point system” to issue green cards to a limited number of unsponsored applicants who have personal attributes associated with successful high-impact, high-tech entrepreneurship, such as graduate education and managerial experience. Such a system would eliminate the linkage between employment and immigration for its beneficiaries, freeing them to engage immediately in entrepreneurial ventures.

The comprehensive immigration bill considered (but not passed) by the U.S. Senate in 2007 contained a point system. The proposed system would have used three selection factors (educational attainment, employment experience, and language proficiency) that are associated with successful entrepreneurship, and two (knowledge of civics and family relationships) that are not known to be. While such a system might be tailored somewhat by dropping the latter factors, it would be very difficult to target potential entrepreneurs precisely. Entrepreneurship is a rare pursuit, even among those with the requisite background for success, and some relevant personal attributes, such as ambition and risk acceptance, are not measurable.

U.S. immigration law currently allocates up to 10,000 green cards annually to applicants who invest $1 million and employ ten U.S. workers. Senators Lugar and Kerry offered a bill in 2010 that would have expanded this “EB-5” category by giving green cards to entrepreneurs who could show ready access to substantial venture capital. These approaches are much more tightly targeted than any point system could be, but their scale is likely to be quite small. Less than 1000 immigrants per year received green cards under EB-5 between 2000 and 2009. Less than 3 percent of the high-impact, high-tech companies in our sample ever received any venture capital investment.

Toward a Long-Term High-Impact Entrepreneurship Strategy

Whether the nation pursues one or more of these three options, or an alternative approach to expand high-impact, high-tech immigrant entrepreneurship, such a step should be seen as one component of a broader strategy to expand high-impact entrepreneurship as a whole. Effective policies for education, research, antitrust, and a variety of other elements of the entrepreneurship environment ought to complement a skills-oriented immigration policy. Such a strategy will lay the groundwork for a sustainable, long-term economic revival.

References are available in the downloadable paper version.

[i] Indeed, the deep roots of these entrepreneurs justify our use of the term “immigrants” in this paper, rather than “foreign-born.” Legally, a person does not “immigrate” to the United States until she obtains a green card.
[ii] Hong Kong and Taiwan are counted in the text and in the table as “countries of origin.”
[iii] The federally-funded New Immigrant Survey finds that “the process of applying for an LPR visa is sufficiently arduous that approximately 22.4 percent of adjustee employment principals became depressed as a result…” (Jasso et al. 2010).
[iv] Some student visa holders are eligible to stay for an additional year after graduation if they are employed in “optional practical training” (OPT) directly related to their field of study. OPT was recently extended to 29 months for graduates in STEM fields.

Downloads

Download the Full Paper

Authors

Zoltan J. Acs
David M. Hart

Image Source: Comstock

The Internet Goes EFL (English as a Foreign Language)

Richard Rose — Wed, 26 Jan 2011 13:51:00 -0500

EXECUTIVE SUMMARY

Communication is central to the exercise of international influence and talking to foreigners is better than fighting them. Communication across national boundaries first of all requires a technology that can span great distances. The Internet and related forms of telecommunication make global conversations far easier, quicker and cheaper than ever before. Secondly, the spread of English as a Foreign Language (EFL) meets the need for a common language and makes it possible for Americans to use their home language when travelling abroad, because others will work in a foreign language in their home country. The third requirement—mutual understanding—is more difficult to achieve. In learning English, people also learn about America, but Americans have fewer incentives and opportunities to understand where foreigners are coming from. Foreigners have not learned EFL to advance America's interests, but to advance their own personal or national interests. If different national interests coincide, then all sides can benefit. But when differences arise, those with a better understanding of others they deal with are better placed to get what they want. The result of this asymmetry is that the international spread of English is increasing the soft power of foreigners who know what they want from Washington—and weakens America's influence with friends as well as potential enemies.

Downloads

Download the Full Paper

Authors

Richard Rose