Nathan Goldschlag
Director of Research, Economic Innovation Group
My research focuses on applied microeconomics of business dynamics, innovation, artificial intelligence, knowledge diffusion, R&D, and entrepreneurship.
AI Charts
Business Dynamism Charts
Research
-
Documents the shift of AI talent from academia to the private sector, driven in part by enormous wage top-end wage premia.
-
Documents how the Census Bureau measures business AI use, synthesizing early evidence across surveys and administrative data.
-
Patents per R&D dollar and growth per patent are not declining; slower growth reflects factors beyond ideas.
-
Non-founder early employees are pivotal to startup success, and losing one has long-run negative effects.
-
Introduces new public data showing the dynamics of high-growth and stagnating firms.
-
Demonstrates that published tables enabled large-scale reconstruction of confidential 2010 Census records.
-
Creates linked microdata to document who inventors are, where they work, and what they earn.
-
AI adoption is reshaping tasks far more than it is reducing employment.
-
Real-time survey data show rapid but still limited AI diffusion with modest employment effects.
-
Large firms adopt automation because they are already large and growing, not because adoption causes growth.
-
No disclosure risk framework satisfies all desiderata, but counterfactual approaches perform well.
-
Trademarks raise variety, quality, and welfare, while also increasing market concentration.
-
Inventors are shifting to large incumbents, earning more but producing fewer innovations.
-
Robot adoption is highly concentrated geographically in a small set of manufacturing hubs.
-
The Characteristics and Geographic Distribution of Robot Hubs in U.S. Manufacturing Establishments 2023Robotics intensity is driven by establishment size and clustered in specialized local ecosystems.
-
Automation is uneven across firms and accounts for 20-30% of the higher labor productivity of large firms in an industry.
-
Details the redesign of the LBD, improving measurement of U.S. firm entry, exit, and growth.
-
High-R&D human capital individuals select into riskier startups with "up-or-out" dynamics.
-
Trademark filing predicts subsequent firm growth and complements patent-based innovation measures.
-
Frontier technologies are rare but concentrated in large firms, amplifying worker exposure.
-
Introduces new BDS tables for High Tech industries, showing their outsized role in job creation.
-
Builds patent-industry concordances to better measure technological change across sectors.
-
Research spending is locally concentrated and deepens firm-university linkages.
-
Introduces scalable machine-learning tools for high-quality business record linkage.
-
Regulatory changes do not explain the broad-based decline in business dynamism.
-
Uses administrative tax data to profile pass-through business owners and their firms.
-
Documents the flow of Ph.D.s on federally funded grants to high-wage industry roles, especially in High Tech sectors.
-
Declining churn has many faces, some less worrisome than others.
Other Writing
- AI and the Real Bottlenecks to Growth
- How Many Businesses Are Using AI?
- Who Are the Inventors and Firms That Generate the Most Patents?
- Peeling the AI Anxiety Onion
- Fat Bear Week and the Fate of the World
- US statistical agencies deserve support and funding. They also need to reform
- AI and Jobs: The Final Word (Until the Next One)
- AI: To Steer or Not To Steer
- High-Tech Industries Are Concentrated in Five Coastal Metro Areas
Datasets & Resources
-
Data and code (with crosswalks) for CPS AI and unemployment CPS AI and unemployment analysis.
-
Public-use data on high growth firm dynamics from the U.S. Census Bureau.
-
Public-use data on business dynamics in high-tech industries.
-
Patent Concordances CrosswalksAlgorithmic patent concordances mapping patent classifications to industry codes. Data files: CPC v1602, CPC v1603, USPC v1601.
-
Replication data and code for research on regulation and business dynamism.
-
Code for classifying industries by STEM occupation concentration.