-
The Rising Returns to R&D: Ideas Are Not Getting Harder to Find
May 2025
Working Paper Number:
CES-25-29
R&D investment has grown robustly, yet aggregate productivity growth has stagnated. Is this because 'ideas are getting harder to find'? This paper uses micro-data from the US Census Bureau to explore the relationship between R&D and productivity in the manufacturing sector from 1976 to 2018. We find that both the elasticity of output (TFP) with respect to R&D and the marginal returns to R&D have risen sharply. Exploring factors affecting returns, we conclude that R&D obsolescence rates must have risen. Using a novel estimation approach, we find consistent evidence of sharply rising technological rivalry. These findings suggest that R&D has become more effective at finding productivity-enhancing ideas but these ideas may also render rivals' technologies obsolete, making innovations more transient.
View Full
Paper PDF
-
The Intangible Divide: Why Do So Few Firms Invest in Innovation?
February 2025
Working Paper Number:
CES-25-15
Investments in software, R&D, and advertising have surged, nearing half of U.S. private nonresidential investment. Yet just a few hundred firms dominate this growth. Most firms, including large ones, regularly invest little in capitalized software and R&D, widening this 'intangible divide' despite falling intangible prices. Using comprehensive US Census microdata, we document these patterns and explore factors associated with intangible investment. We find that firms invest significantly less in innovation-related intangibles when their rivals invest more. One firm's investment can obsolesce rivals' investments, reducing returns. This negative pecuniary externality worsens the intangible divide, potentially leading to significant misallocation.
View Full
Paper PDF
-
U.S. Banks' Artificial Intelligence and Small Business Lending: Evidence from the Census Bureau's Annual Business Survey
February 2025
Working Paper Number:
CES-25-07
Utilizing confidential microdata from the Census Bureau's new technology survey (technology module of the Annual Business Survey), we shed light on U.S. banks' use of artificial intelligence (AI) and its effect on their small business lending. We find that the percentage of banks using AI increases from 14% in 2017 to 43% in 2019. Linking banks' AI use to their small business lending, we find that banks with greater AI usage lend significantly more to distant borrowers, about whom they have less soft information. Using an instrumental variable based on banks' proximity to AI vendors, we show that AI's effect is likely causal. In contrast, we do not find similar effects for cloud systems, other types of software, or hardware surveyed by Census, highlighting AI's uniqueness. Moreover, AI's effect on distant lending is more pronounced in poorer areas and areas with less bank presence. Last, we find that banks with greater AI usage experience lower default rates among distant borrowers and charge these borrowers lower interest rates, suggesting that AI helps banks identify creditworthy borrowers at loan origination. Overall, our evidence suggests that AI helps banks reduce information asymmetry with borrowers, thereby enabling them to extend credit over greater distances.
View Full
Paper PDF
-
Financing, Ownership, and Performance: A Novel, Longitudinal Firm-Level Database
December 2024
Working Paper Number:
CES-24-73
The Census Bureau's Longitudinal Business Database (LBD) underpins many studies of firm-level behavior. It tracks longitudinally all employers in the nonfarm private sector but lacks information about business financing and owner characteristics. We address this shortcoming by linking LBD observations to firm-level data drawn from several large Census Bureau surveys. The resulting Longitudinal Employer, Owner, and Financing (LEOF) database contains more than 3 million observations at the firm-year level with information about start-up financing, current financing, owner demographics, ownership structure, profitability, and owner aspirations ' all linked to annual firm-level employment data since the firm hired its first employee. Using the LEOF database, we document trends in owner demographics and financing patterns and investigate how these business characteristics relate to firm-level employment outcomes.
View Full
Paper PDF
-
Multinational Production and Innovation in Tandem
October 2024
Working Paper Number:
CES-24-64
Multinational firms colocate production and innovation by offshoring them to the same host country or region. In this paper, I examine the determinants of multinational firms' production and innovation locations. Exploiting plausibly exogenous variations in tariffs, I find complementarities between production and innovation within host countries and regions. To evaluate manufacturing reshoring policies, I develop a quantitative multicountry offshoring location choice model. I allow for rich colocation benefits and cross-country interdependencies and prove supermodularity of the model to solve this otherwise NP-hard problem. I find the effects of manufacturing reshoring policies are nonlinear, contingent upon firm heterogeneity, and they accumulate dynamically.
View Full
Paper PDF
-
Exploratory Report: Annual Business Survey Ownership Diversity and Its Association with Patenting and Venture Capital Success
October 2024
Working Paper Number:
CES-24-62
The Annual Business Survey (ABS) as the replacement for the Survey of Business Owners (SBO) serves as the principal data source for investigating business ownership of minorities, women, and immigrants. As a combination of SBO, the innovation questions formerly collected in the Business R&D and Innovation Survey (BRDIS), and an R&D module for microbusinesses with fewer than 10 employees, ABS opens new research opportunities investigating how ownership demographics are associated with innovation. One critical issue that ABS is uniquely able to investigate is the role that diversity among ownership teams plays in facilitating innovation or intermediate innovation outcomes in R&D-performing microbusinesses. Earlier research using ABS identified both demographic and disciplinary diversity as strong correlates to new-to-market innovation. This research investigates the extent to which the various forms of diversity also impact tangible innovation related intermediate outcomes such as the awarding of patents or securing venture capital financing for R&D. The other major difference with the earlier work is the focus on R&D-performing microbusinesses that are an essential input to radical innovation through the division of innovative labor. Evidence that disciplinary and/or demographic diversity affect the likelihood of receiving a patent or securing venture capital financing by small, high-tech start-ups may have implications for higher education, affirmative action, and immigration policy.
View Full
Paper PDF
-
Grassroots Design Meets Grassroots Innovation: Rural Design Orientation and Firm Performance
March 2024
Working Paper Number:
CES-24-17
The study of grassroots design'applying structured, creative processes to the usability or aesthetics of a product without input from professional design consultancies'remains under investigated. If design comprises a mediation between people and technology whereby technologies are made more accessible or more likely to delight, then the process by which new grassroots inventions are transformed into innovations valued in markets cannot be fully understood. This paper uses U.S. data on the design orientation of respondents in the 2014 Rural Establishment Innovation Survey linked to longitudinal data on the same firms to examine the association between design, innovation, and employment and payroll growth. Findings from the research will inform questions to be investigated in the recently collected 2022 Annual Business Survey (ABS) that for the first time contains a Design module.
View Full
Paper PDF
-
Tracking Firm Use of AI in Real Time: A Snapshot from the Business Trends and Outlook Survey
March 2024
Working Paper Number:
CES-24-16R
Timely and accurate measurement of AI use by firms is both challenging and crucial for understanding the impacts of AI on the U.S. economy. We provide new, real-time estimates of current and expected future use of AI for business purposes based on the Business Trends and Outlook Survey for September 2023 to February 2024. During this period, bi-weekly estimates of AI use rate rose from 3.7% to 5.4%, with an expected rate of about 6.6% by early Fall 2024. The fraction of workers at businesses that use AI is higher, especially for large businesses and in the Information sector. AI use is higher in large firms but the relationship between AI use and firm size is non-monotonic. In contrast, AI use is higher in young firms. Common uses of AI include marketing automation, virtual agents, and data/text analytics. AI users often utilize AI to substitute for worker tasks and equipment/software, but few report reductions in employment due to AI use. Many firms undergo organizational changes to accommodate AI, particularly by training staff, developing new workflows, and purchasing cloud services/storage. AI users also exhibit better overall performance and higher incidence of employment expansion compared to other businesses. The most common reason for non-adoption is the inapplicability of AI to the business.
View Full
Paper PDF
-
AI Adoption in America: Who, What, and Where
September 2023
Working Paper Number:
CES-23-48R
We study the early adoption and diffusion of five AI-related technologies (automated-guided vehicles, machine learning, machine vision, natural language processing, and voice recognition) as documented in the 2018 Annual Business Survey of 850,000 firms across the United States. We find that fewer than 6% of firms used any of the AI-related technologies we measure, though most very large firms reported at least some AI use. Weighted by employment, average adoption was just over 18%. AI use in production, while varying considerably by industry, nevertheless was found in every sector of the economy and clustered with emerging technologies such as cloud computing and robotics. Among dynamic young firms, AI use was highest alongside more educated, more-experienced, and younger owners, including owners motivated by bringing new ideas to market or helping the community. AI adoption was also more common alongside indicators of high-growth entrepreneurship, including venture capital funding, recent product and process innovation, and growth-oriented business strategies. Early adoption was far from evenly distributed: a handful of 'superstar' cities and emerging hubs led startups' adoption of AI. These patterns of early AI use foreshadow economic and social impacts far beyond this limited initial diffusion, with the possibility of a growing 'AI divide' if early patterns persist.
View Full
Paper PDF
-
Registered Report: Exploratory Analysis of Ownership Diversity and Innovation in the Annual Business Survey
March 2023
Working Paper Number:
CES-23-11
A lack of transparency in specification testing is a major contributor to the replicability crisis that has eroded the credibility of findings for informing policy. How diversity is associated with outcomes of interest is particularly susceptible to the production of nonreplicable findings given the very large number of alternative measures applied to several policy relevant attributes such as race, ethnicity, gender, or foreign-born status. The very large number of alternative measures substantially increases the probability of false discovery where nominally significant parameter estimates'selected through numerous though unreported specification tests'may not be representative of true associations in the population. The purpose of this registered report is to: 1) select a single measure of ownership diversity that satisfies explicit, requisite axioms; 2) split the Annual Business Survey (ABS) into an exploratory sample (35%) used in this analysis and a confirmatory sample (65%) that will be accessed only after the publication of this report; 3) regress self-reported new-to-market innovation on the diversity measure along with industry and firm-size controls; 4) pass through those variables meeting precision and magnitude criteria for hypothesis testing using the confirmatory sample; and 5) document the full set of hypotheses to be tested in the final analysis along with a discussion of the false discovery and family-wise error rate corrections to be applied. The discussion concludes with the added value of implementing split sample designs within the Federal Statistical Research Data Center system where access to data is strictly controlled.
View Full
Paper PDF