Measuring and Accounting for Innovation in the 21st Century

March 10-11, 2017
Carol Corrado, the Conference Board, Javier Miranda, Bureau of the Census, Jonathan Haskel, Imperial College London, and Daniel Sichel, Wellesley College, Organizers

Extending Current Measurement Frameworks

Charles Hulten, the University of Maryland and NBER, and Leonard Nakamura, Federal Reserve Bank of Philadelphia

We See the Digital Revolution Everywhere Except in Real GDP

Katharine Abraham and John Haltiwanger, the University of Maryland and NBER, and Kristin Sandusky and James Spletzer, Bureau of the Census

Measuring the Gig Economy: Current Knowledge and Open Issues (NBER Working Paper No. 24950)

The rise of the "gig economy" has attracted wide attention from both scholars and the popular media. Much of this attention has been devoted to jobs mediated through various online platforms. While non-traditional work arrangements have been a perennial subject of debate and study, the perception that new technology is producing an accelerated pace of change in the organization of work has fueled a resurgence of interest in how such changes may be affecting both workers and firms. In this paper, Abraham, Haltiwanger, Sandusky, and Spletzer provides a typology of work arrangements and reviews how different arrangements, and especially gig activity, are captured in existing data. A challenge for understanding recent trends is that household survey and administrative data paint a different picture, with the former showing little evidence of the growth in self-employment that would be implied by a surge in gig activity and the latter providing evidence of considerable recent growth. An examination of matched individual-level survey and administrative records shows that a large and growing fraction of those with self-employment activity in administrative data have no such activity recorded in household survey data. The share of those with self-employment activity in household survey data but not administrative data is smaller and has not grown. Promising avenues for improving the measurement of self-employment activity include the addition of more probing questions to household survey questionnaires and the development of integrated data sets that combine survey, administrative and, potentially, private data.

Dominique Guellec, OECD/OCDE, and Caroline Paunov, OECD

Digital Innovation and the Distribution of Income (NBER Working Paper No. 23987)

Income inequalities have increased in most OECD countries over the past two to three decades; particularly the income share of the top 1% has soared. In this paper Paunov and Guellec argue that the increasing importance of digital innovation — new products and processes based on software code and data — has increased market rents, which benefit disproportionately the top income groups. In line with a Schumpeterian vision, innovation gives rise to rents from market power and scale economies. This is magnified with digital innovation, in which the intangible component (the source of rents) is much larger than in traditional manufacturing innovation. Highly concentrated market structures ("winner-take-all") allow rent extraction. In addition, digital innovation tends to increase risks because even only marginally superior products can take over the entire market, hence rendering market shares unstable. Instability commands risk premia, hence higher expected revenues, for investors. Market rents accrue mainly to investors and top managers and less to the average workers, hence increasing income inequality. Market rents are needed to incentivize innovation and compensate for its costs, but beyond a certain level they become detrimental as rent seeking then substitutes to innovation in business strategies. Public policy may stimulate innovation and welfare by eliminating ex ante the market conditions which allow rent extraction that comes from anti-competitive practices.

Lucia Foster, Cheryl Grim, and Zoltan Wolf, Bureau of the Census, and John Haltiwanger, University of Maryland and NBER

Innovation, Productivity Dispersion, and Productivity Growth
Chapter in the NBER book Measuring and Accounting for Innovation in the 21st Century

Foster, Grim, Haltiwanger, and Wolf examine whether underlying industry innovation dynamics are an important driver of the large dispersion in productivity across firms within narrowly defined sectors. Their hypothesis is that periods of rapid innovation are accompanied by high rates of entry, significant experimentation and, in turn, a high degree of productivity dispersion. Following this experimentation phase, successful innovators and adopters grow while unsuccessful innovators contract and exit yielding productivity growth. The researchers examine the dynamic relationship between entry, productivity dispersion, and productivity growth using a new comprehensive firm-level dataset for the U.S. Foster, Grim, Haltiwanger, and Wolf find a surge of entry within an industry yields with a lag an increase in productivity dispersion and then after a subsequent lag an increase in productivity growth. These patterns are more pronounced for the High Tech sector where we expect there to be more innovative activities. These patterns change over time suggesting other forces are at work during the post-2000 slowdown in aggregate productivity.

New Approaches and Data

Wesley Cohen, Duke University and NBER, and You-Na Lee and John Walsh, Georgia Institute of Technology

Measuring the Several Faces of Innovation

Emin Dinlersoz, Nathan Goldschlag, and Nikolas Zolas, Bureau of the Census, and Amanda Myers, United States Patent and Trademark Office

An Anatomy of U.S. Firms Seeking Trademark Registration (NBER Working Paper No. 25038)

Dinlersoz, Goldschlag, Myers, and Zolas report on the construction of a new dataset that combines data on trademarks from the U.S. Patent and Trademark Office with the microdata on firms from the U.S. Census Bureau. The methodology for merging the data and identifying matches between trademarks and firms is described. The resulting dataset allows tracking of various activity related to trademarks over the life-cycle of firms, such as the first application for a trademark, the first use of a trademark, and the renewal, assignment, and abandonment of trademarks. Some facts about firm-level trademarking activity are documented, including the incidence and timing of trademarking activity over the firm life-cycle, the connection between firm characteristics and trademarking, and the relation of trademarks to firm growth. The dataset offers new possibilities for research on how trademarks are related to firm dynamics and performance, firm-level innovative activity and product introductions, and firm strategies aimed at acquiring customers, generating loyalty, building brands and reputation, and signalling quality.

Nathan Goldschlag, Ron Jarmin, and Nikolas Zolas, Bureau of the Census, and Julia Lane, New York University

The Link between University R&D, Human Capital and Business Startups

Goldschlag, Jarmin, Zolas, and Lane expand the data infrastructure available to build evidence on public and private investments in science and R&D and utilize it to examine the links between startup performance and new measures of workforce human capital. They apply machine-learning techniques to a rich new source of longitudinally-linked data to characterize the research-experienced workforce of new businesses. Startups with a more research-experienced workforce are more likely to survive and grow.

Javier Miranda and Nikolas Zolas, Bureau of the Census

Measuring the Impact of Household Innovation Using Non Employer Administrative Data

Miranda and Zolas link USPTO patent data to U.S. Census Bureau administrative records on individuals and firms. The combined dataset provides a directory of patenting household inventors as well as a time-series directory of self-employed businesses tied to household innovations. The researchers describe the characteristics of household inventors by race, age, gender and U.S. origin, as well as the types of patented innovations pursued by these inventors. Business data allows them to highlight how patents shape the early life-cycle dynamics of nonemployer businesses. The researchers find household innovators are disproportionately U.S. born, white and older relative to business innovators. Data shows there is a deficit of female and black inventors. Household inventors tend to work in consumer product areas compared to traditional business patents. While patented household innovations do not have the same impact of business innovations their uniqueness and impact remains surprisingly high. Back of the envelope calculations suggest patented household innovations might generate between $7.2B and $8.2B in revenue.

Improving Current Measurement Frameworks

Michael Polder, Statistics Netherlands; Pierre Mohnen, Maastricht University; and George van Leeuwen, Statistics Netherlands

ICT, R&D and Organizational Innovation: Exploring Complementarities in Investment and Production (NBER Working Paper No. 25044)

Polder, Mohnen, and van Leeuwen investigate with Dutch micro data whether ICT investment boosts total factor productivity, whether it does so because it increases the return to R&D, and whether ICT requires organizational innovations to increase productivity.

Wen Chen, Bart Los, and Marcel Timmer, University of Groningen

Measuring the Returns to Intangibles: a Global Value Chain Approach

Considerable progress has been made in tracing expenditures on intangibles in the macroeconomy. But much less is known about their returns. In this paper Los, de Vries, and Timmer outline a new strategy to estimate returns to intangibles in the context of globalised production networks. The researchers view intangibles as inputs that allow a firm to generate surplus value from tangible factor inputs. This is in contrast to the standard treatment of intangible capital as yet another factor of production which can be separately valued. The researchers propose an instrumental definition of the returns to intangibles as the residual value after subtracting the costs of labour and tangible capital. Given international fragmentation of production processes, this residual can only be measured when all stages of production (including distribution) are considered. To this end, they rely on the global value chain (GVC) approach introduced by Timmer et al. (2014). This approach allows for a decomposition of the value of a product into value added at each stage of production. The researchers extend this approach by splitting value added into returns to labour, tangible and intangible capital. They focus on final manufacturing products for the period 1995-2007 using the world input-output database (WIOD) and additional data derived from national accounts statistics. The researchers' main finding is that the share of intangibles in the value of final products has increased from 2000 onwards. This is found for all manufacturing product groups. They also find that for buyer-driven GVCs (like food, textiles and furniture) returns to intangibles are increasingly realized in the distribution stage, that is, in delivery of the final product to the consumer. In contrast, in producer-driven GVCs (like machinery, automotive and electronics) the returns are shifting to activities before the final production stage.

Alexis N. Grimm, Bureau of Economic Analysis

Trends in U.S. Trade in Information and Communications Technology (ICT) Services and in ICT-Enabled Services

Kenneth Flamm, University of Texas at Austin

Has Moore's Law Been Suspended or Repealed? An Empirical Economic Analysis of the Pace of Innovation in Semiconductors
Chapter in the NBER book Measuring and Accounting for Innovation in the 21st Century

Carol Corrado, and David Byrne, Federal Reserve Board

Accounting for Innovation in Consumer ICT Services

Previous work (Byrne and Corrado, 2017a,b) reassessed the link between ICT investment prices and productivity to help understand how digital technology contributes to labor productivity growth. This paper extends the analysis by considering consumer spending on digital durable goods as investment. When consumer digital stocks are treated as investment, their asset prices have implications for productivity, and the researchers find mismeasurement patterns similar to those found in their earlier work on private ICT investment goods. They also consider whether and how households' increased use of its stocks of digital goods should be folded into the measurement of income from those stocks. Without considering use rates, the researchers find that real services from consumer use of their digital stocks has increased nearly 20 percent per year since 1985. After adjusting for the increase in utilization of those stocks since 1995, real consumption of digital services is estimated to have grown 35 percent per year. The utilization adjustment is especially potent between 2005 and 2012.

David Byrne, Federal Reserve Board; Carol Corrado; and Daniel E. Sichel

The Rise of Cloud Computing: Minding Your P's, Q's and K's (NBER Working Paper No. 25188)

Cloud computing—computing done on an off-site network of resources accessed through the Internet—is revolutionizing how computing services are used. However, because cloud is so new and it largely is an intermediate input to other industries, it is difficult to track in the U.S. statistical system. Moreover, there is a paucity of systematic information on the prices of cloud services. To begin filling this gap, this paper does three things. First, Byrne, Corrado, and Sichel define the different segments of cloud computing and document its explosive expansion. Second, they develop new hedonic prices indexes for cloud services based on quarterly data for compute, database, and storage services offered by Amazon Web Services (AWS) from 2009 to 2016. These indexes fall rapidly over the sample period, with quickening (and double digit) rates of decline for all three products starting at the beginning of 2014. Finally, the researchers highlight the puzzle of why investment in IT equipment in the NIPAs has been so weak while capital expenditures have exploded for IT equipment associated with cloud infrastructure. Byrne, Corrado, and Sichel suggest that cloud service providers are undertaking large amounts of own-account investment in IT equipment and that some of this investment may not be captured in GDP.

Ana Aizcorbe and David Wasshausen, Bureau of Economic Analysis

BEA Deflators for Information and Communications Technology Goods and Services Historical Analyses and Future Plans

NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us