2020-2021 Seminar Series

Thank you to all the presenters and attendees for our 2020-2021 seminar series.

            2020-2021 talks:

            • October 7: Joowon Kim, “Political Spending for Firm Performance: Evidence from Citizens United”
            • October 21: Anna Salomons, “New Frontiers: The Origins and Evolving Content of New Work”

            Abstract: Recent theory stresses the importance of new job types (‘new work’) that emerge as automation subsumes existing labor tasks. Comprehensive and representative empirical evidence on this phenomenon is lacking, however. We construct an inventory of new job titles linked to United States Census microdata over 1940-2018 and explore the emergence and evolution of new work—including its skill demands and wage levels. Comparing new to pre-existing (‘old’) work, our descriptive analysis detects sharp shifts across decades in the content of new work and the education and wage levels of those who perform it. In the first four post-War decades, new work emerges disproportionately in production and office occupations, and in professional specialties. Conversely, new work creation since 1980 has increasingly concentrated in high-education specialties and, to a lesser extent, in low-education personal services and manual occupations. In both eras, new work emerges disproportionately in activities where old work is also growing. Leveraging these observations, we offer a high-level hypothesis for the sources of new work creation and present evidence on empirical tests of this hypothesis.

            • November 4: Enjar Lkhagvajav, “Patent Disclosure, Firm Innovation, and Growth”

            Abstract: A patent system requires innovators to reveal their new ideas in return for monopoly rights to their uses. This paper shows that patent disclosure requirements can discourage firm patenting both empirically and theoretically. In the data, I analyze the effect of the American Inventor’s Protection Act of 1999 (AIPA), which shortened publication time for patents filed after 2000. Due to earlier patent disclosure, U.S. public firms lowered their patenting and R&D growth. I then build a Schumpeterian endogenous growth model with firm innovation and patenting, together with a role for disclosure policy. At the firm level, the model reveals that patent disclosure can lower patenting due to a trade-off between costly disclosure and patent monopoly protection. The model matches the empirical evidence of lower firm patenting due to higher patent disclosure. At the macro level, the patent disclosure could reduce patenting and overall economic growth, contributing to recent U.S. trends.

            Abstract: Is a coherent view of digital policy possible? The U.S. has not attempted a top-level approach to digital policy since the Clinton administration, but the global rise of American digital giants has inspired efforts around the world to understand digital economics, especially as it affects competition. The pandemic has elevated the value of digital services, compelling attention to the benefits – and to concerns about scale and concentration. At the same time, digital economics is increasingly entangled with neighboring domains – privacy, intellectual property, tax, research funding, harmful content, trade, industrial policy, and national security. It colors macro-issues that lack an institutional home, such as inequality, dynamism, and the future of work. And it is driving policy divergence internationally – notably between the U.S., Europe, and China – along with developments in techno-nationalism and geo-economics.

            This presentation builds on work for OECD’s Going Digital project, specifically Vectors of Digital Transformation, which, while scrupulously apolitical, presented a stylized framework for understanding the economic dimensions of digitization. Connecting evolving real-world experience with digital economics is challenging enough under more stable conditions. Under present circumstances, how can deeper understanding of digital phenomena and dynamics be synthesized for the benefit of policymakers?

            • December 2: Michael Impink, “Barriers to Growth in Developing AI”

            Abstract: Artificial Intelligence products are expected to increase labor productivity and drive macroeconomic growth. For these gains to be realized, AI startups must be able to raise the funding needed to develop their underlying AI technologies and resulting products. Training data is important to startups developing AI; however, there is no consensus in the literature on aspects of training data that are most important to acquire needed funding. In this paper, we explore if startups that lack proprietary training from suppliers can acquire the funding needed to grow. We develop a framework for characterizing training data as a resource derived from relationships with customers and suppliers, and then use this framework to describe conditions under which such data can lead to competitive advantage. Using unique data from two waves of surveys of AI startups, we find that without additional training data from cloud services providers, above and beyond access to proprietary customer data, AI startups may not be able to acquire the funding needed to grow.

            Abstract: This paper investigates the impact of investment in automation- and AI- related goods on within-firm wage inequality in the French economy during the period 2002-2017. We document that most of wage inequality in France is accounted for by differences among workers belonging to the same firm, rather than by differences between sectors, firms, and occupations. Using an event-study approach on a sample of firms importing automation and AI-related goods, we find that automation/AI spikes are not followed by an increase in within-firm wage and gender inequality. Instead, wages tend to increase at different percentiles of the distribution, revealing a relatively spread allocation of rents from automation/AI within the firm. This adds to previous findings showing picture of a `labor friendly’ effect of the latest wave of new technologies.

            • February 17: James Bessen, “Firm Differences: Skill Sorting and Software” (hosted by Utrecht University)

            Abstract: Research shows that much of the recent rise in wage inequality comes from growing differences between firms, especially sorting of skilled workers to high-paying firms. This paper explores the role of proprietary software in these changes. Using job ad data, we find that proprietary software is strongly associated with firm wage fixed effects and also with firm skills. Software accounts for half or more of skill sorting across firms. Moreover, the role of software in sorting is greater at larger firms. Large investment in proprietary software helps explain the growth in skill sorting that increases between-firm wage inequality.

            Abstract: This paper uses detailed panel data at the buyer level to understand the size and nature of spillovers across product categories in medical devices. Difference-in-differences regressions identify spillovers in usage (accounting for 21 percent of market share) across categories that share physical features. All categories exhibit spillovers in contracting (accounting for 7 to 18 percent of the probability of contracting). These buyer-level spillovers represent up to half of all potential demand spillovers and one-third of the overall correlation in shares across categories, suggesting meaningful implications for firm strategy and antitrust policy for firms selling in multiple product categories.

            • April 7: Nathaniel Breg, “Is Robotic Surgery Good for Patients? Evidence from Hysterectomy”

            Abstract: Robotically-assisted soft-tissue surgery has been increasingly prevalent in U.S. health care, yet there is little rigorous evidence as to its efficacy relative to alternative approaches, including laparoscopic surgery and open surgery. This paper analyzes Medicare inpatient claims from 2008 – 2012 for a common procedure, total hysterectomy, and estimate the effect on patient outcomes of robotic surgery relative to other modes of surgery. The estimation controls for a wealth of patient characteristics and for hospital and physician fixed effects. I implement two alternative instrumental variables strategies exploiting different sources of variation. In one strategy, I instrument for the robotic treatment choice using an indicator for whether a hospital referral region has robotic surgery. In the other strategy, the instrument is a measure of physician preference for the robotic approach. Across specifications, I find that robotic surgery lowers the probability of a patient staying in the hospital for two or more days by 47 to 70 percentage points (off a mean of 94%). The robotic approach also lowers the percent of patients experiencing an all-cause 90-day readmission by 0 to 11 percentage points (off a mean of 16%) relative to all other approaches. Descriptive evidence suggests that most of the robotic approach’s substantial effect on outcomes in the population comes from its large advantage over open surgery among patients who would have had open surgery if the robotic option were not available to them.

            • April 21: Cäcilia Lipowski, “Technological Change and Labor Market Opportunities”

            Abstract: The wage penalty of a disadvantaged family background declined in the past decades in Germany, even though wage inequality rose. The role of skill-biased technological change for overall wage inequality is well documented. This paper shows that technological progress can also explain the decline of the wage penalty of a disadvantaged family background in Germany between 1986 and 2012, and the rising share of workers with disadvantaged parental background in high-wage jobs. Intuitively, technological change increases returns to individual ability relative to the returns to parental background, as parents’ occupation specific knowledge and networks depreciate during rapid technological transformations.

              • May 5: Fabio Landini, “Robots, Digitalization, and Worker Voice.”

              Abstract: The interplay between labour institutions and the adoption of automation technologies remains poorly understood. Specifically, there is little evidence on how the nature of industrial relations shapes technological choices at the workplace level. Using a large sample of more than 20000 European establishments located in 26 countries, this paper documents a positive association between the presence of employee representation (ER) and the use of automation technologies. We account for the endogeneity of ER bodies by using instrumental variables and exploiting size-contingent regulations in the context of a local-randomization regression discontinuity design. We extensively dig into the mechanisms through which ER may foster the use of automation technologies by exploiting rich information on the de facto role played by ER bodies in relation to well-defined decision areas of management. Greater automation in establishments with ER does not seem to be driven by employers’ strategic bargaining considerations in the context of adversarial labour-management relationships (as measured by past strike activity) or constraints on labour flexibility imposed by the interference of employee representatives with dismissal procedures. We find suggestive evidence that ER influences work organization and certain workplace practices, such as training, working time management and information sharing, that may be complementary to new technologies.

              • May 12: James Bessen, “From Productivity to Firm Growth.”

              Abstract: It is widely held that more productive firms grow faster, thus reallocating resources and raising aggregate productivity. Yet little empirical research identifies the features of the mechanisms affecting this process. This paper develops and tests a general model encompassing several mechanisms used to overcome informational frictions to growth. We find that firm size, productivity dispersion, and large firm investments in intangibles are all significantly related to changes in firm growth in response to productivity. These factors can account for much of the decline in the response to productivity since 2000 (Decker et al. 2020). Also, industry concentration is directly related to aggregate productivity growth.

              • May 19: Emiel van Bezooijen, “The Young Bunch: The Impact of Youth Minimum Wages on Youth Labor Market Outcomes.”

              Abstract: We estimate the effects of an increase in the minimum wage for 20- to 22-year-olds in the Netherlands on their employment and earnings, using a difference-in-differences approach with detailed administrative data. We find that the increase did not have a negative effect on the number of jobs and even somewhat increased hours worked, and hence raised overall earnings for all three age-groups. Further, the minimum wage increase has substantial spillover effects, accounting for more than 75% of the average wage increase experienced by workers. Lastly, while employment grows in fixed-term and temporary help agency contracts, we do not find evidence of declines in employment in other types of work arrangements, nor of labor-labor substitution. Importantly, labor market outcomes evolve most favorably for full-time workers who are not enrolled in education and thus less likely to be transient occupants of minimum-wage jobs.

              • May 26: Alexandra Spitz-Oener, “Workers are not machines, they are humans and they have the capacity to adjust!”
              • June 9: Raviv Murciano-Goroff, “Hidden Software and Veiled Value Creation: Illustrations from Server Software Usage”

              Abstract: How do you measure the value of a commodity that transacts at a price of zero from an economic standpoint? This study examines the potential for and extent of omission and misattribution in standard approaches to economic accounting with regards to open source software, an unpriced commodity in the digital economy. The study is the first to follow usage and upgrading of unpriced software over a long period of time. It finds evidence that software updates mislead analyses of sources of firm productivity and identifies several mechanisms that create issues for mismeasurement. To illustrate these mechanisms, this study closely examines one asset that plays a critical role in the digital economic activity, web server software. We analyze the largest dataset ever compiled on web server use in the United States and link it to disaggregated information on over 200,000 medium to large organizations in the United States between 2001 and 2018. In our sample, we find that the omission of economic value created by web server software is substantial and that this omission indicates there is over $4.5 billion dollars of mismeasurement of server software across organizations in the United States. This mismeasurement varies by organization age, geography, industry and size. We also find that dynamic behavior, such as improvements of server technology and entry of new products, further exacerbates economic mismeasurement.

              • June 16: Jennifer Hunt, “Is Distance from Innovation a Barrier to the Adoption of Artificial Intelligence?“

              Abstract: We investigate whether posted vacancies for jobs requiring Artificial Intelligence (AI) skills grow more slowly in U.S. locations farther from AI innovation hotspots. To define hotspots, we create a geocoded dataset of all AI journal publications or conference proceedings through 2020, while we obtain the job vacancy information from online job advertisements scraped by Burning Glass Technologies from 2007–2019. We find that a commuting zone’s AI publications affect other commuting zones’ AI vacancies if it had at least 750 AI publications by 2006, which 5% of commuting zones did. A 10% greater distance from such a hotspot (about a standard deviation) reduces a commuting zone’s 2007–2019 growth in AI jobs’ share of job advertisements by 4–5% of median growth. This suggests that distance from innovation is a barrier to the adoption of technology.