-
AI Exposure and Adoption Among U.S. Firms
September 2026
Authors:
John Haltiwanger,
Lucia Foster,
Martha Stinson,
Emin Dinlersoz,
Cheryl Grim,
Zoltan Wolf,
Sabrina Wulff Pabilonia,
Matthew Dey,
Sean Wang,
Aditya Pande,
Peter B. Meyer
Working Paper Number:
CES-26-61
Measures of exposure to artificial intelligence (AI) are widely used to study where AI is likely to affect firm and worker outcomes, yet limited evidence exists on how closely exposure relates to realized AI adoption at the firm level. We examine this relationship by linking 14 firm-level exposure measures, constructed from occupational exposure estimates in the literature and occupational employment shares from the Bureau of Labor Statistics Occupational Employment and Wage Statistics program, to direct measures of firm AI adoption from the Census Bureau's Business Trends and Outlook Survey. Exposure is positively and statistically significantly associated with adoption, but explains only a modest share of its variation. The strength of the relationship varies considerably across standardized exposure measures: a one-standard-deviation increase in firm-level exposure is associated with a 4'11 percentage point higher firm adoption probability, falling to 1'8 percentage points after controlling for year and sub-sector fixed effects. The exposure'adoption relationship is also heterogeneous across firm-size classes and sectors. Exposure is thus an informative but incomplete signal of adoption. More broadly, the results have implications for the interpretation of exposure-based measures in studies of the economic effects of AI.
View Full
Paper PDF
-
Cleansing Floods? Creative Destruction and Government Spending After Disasters
August 2026
Working Paper Number:
CES-26-52
Disasters devastate economies, yet productivity can improve in their wake. Why? Using confidential plant-level microdata from the US Census Bureau and an event study design, I trace the creative destruction process that follows large federally declared floods. Exits are concentrated among the least productive plants, whose used machines are then acquired by high-productivity entrants. Survivors upgrade their machinery as they rebuild and see productivity gains. Federal disaster spending facilitates this process by expanding financing access for nimble young and small firms that disproportionately fuel productive reallocation. Without it, financing constraints stifle creative destruction and productivity declines. Ultimately, the relative income gains from federal disaster assistance generate tax revenues far exceeding the policy's upfront cost, making it both efficiency-enhancing and fiscally sound. My findings reveal a novel allocative efficiency channel through which government spending supports post-disaster recovery, with critical implications for a warming world.
View Full
Paper PDF
-
Desperate Capital Breeds Productivity Loss: Evidence From Public Pension Investments in Private Equity
July 2026
Working Paper Number:
CES-26-47
I study investments of U.S. public pensions in private equity (PE), and trace them to ultimate micro assets'target firms which PE funds invest in, using micro-data on private investments combined with confidential U.S. Census data. I show that more severely underfunded public pensions receive lower average PE returns, and match with smaller GPs on average, than less underfunded pensions. Consistent with matching and returns, firms financed by most underfunded public pensions and smallest PE funds face labor productivity decreases. I introduce a novel instrument'public unionization'in support of underfunding positions driving selection into funds. Lastly, I evaluate alternate mechanisms.
View Full
Paper PDF
-
The Adoption of Non-Rival Inputs and Firm Scope
April 2026
Working Paper Number:
CES-26-28
Custom software is distinct from other types of capital in that it is non-rival'once a firm makes an investment in custom software, it can be used simultaneously across its many establishments. Using confidential U.S. Census data, we document that while firms with more establishments are more likely to invest in custom software, they spend less on it as a share of total capital expenditure. We explain these empirical patterns by developing a model that incorporates the non-rivalry of custom software. In the model, firms choose whether to adopt custom software, the intensity of their investment, and their scope, balancing the cost of managing multiple establishments with the increasing returns to scope from the nonrivalrous custom software investment. Using the calibrated model, we assess the extent to which the decline in the rental rate of custom software over the past 40 years can account for a number of macroeconomic trends, including increases in firm scope and concentration.
View Full
Paper PDF
-
Allocating Misallocation: Decomposing Measures of Aggregate Allocative Efficiency
April 2026
Working Paper Number:
CES-26-26
We explore sources of measured misallocation using establishment data from U.S. manufacturing industries. We decompose standard revenue productivity dispersion statistics into contributions by dispersion in revenue margins over costs and dispersion in input cost shares across plants. We establish a formal link between these components and measured allocative efficiency. The results indicate the components contribute similarly to apparent rising misallocation in US manufacturing. We use the mapping between distortions that influence these distinct components to explore the relationship between inferred distortions and mechanisms that influence one or both sources of revenue productivity dispersion. Finally, we show rising misallocation in the US manufacturing sector in the last several decades is pervasive, and yet a few industries account for over half of the aggregate decline.
View Full
Paper PDF
-
Specialization in a Knowledge Economy
December 2025
Working Paper Number:
CES-25-77
Using firm-level data from the US Census Longitudinal Business Database (LBD), this paper exhibits novel evidence about a wave of specialization experienced by US firms in the 1980s and 1990s. Specifically: (i) Firms, especially innovating ones, decreased production scope, i.e., the number of industries in which they produce. (ii) Innovation and production separated, with small firms specializing in innovation and large firms in production. Higher patent trading efficiency and stronger patent protection are proposed to explain these phenomena. An endogenous growth model is developed with potential mismatches between innovation and production. Calibrating the model suggests that increased trading efficiency and better patent protection can explain 20% of the observed production scope decrease and 108% of the innovation and production separation. They result in a 0.64 percent point increase in the annual economic growth rate. Empirical analyses provide evidence of causality from pro-patent reforms in the 1980s to the two specialization patterns.
View Full
Paper PDF
-
Borrowing Constraints, Markups, and Misallocation
December 2025
Working Paper Number:
CES-25-75
We document new facts that link firms' markups to borrowing constraints: (1) less constrained firms within an industry have higher markups, especially in industries where assets are difficult to borrow against and firms rely more on earnings to borrow; (2) markup dispersion is also higher in industries where firms rely more on earnings to borrow. We explain these relationships using a standard Kimball demand model augmented with borrowing against assets and earnings. The key mechanism is a two-way feedback between markups and borrowing constraints. First, less constrained firms charge higher markups, as looser constraints allow them to attain larger market shares. Second, higher markups relax borrowing constraints when firms rely on earnings to borrow, as those with higher markups have higher earnings. This two-way feedback lowers TFP losses from markup dispersion, particularly when firms rely on earnings to borrow.
View Full
Paper PDF
-
Technifying Ventures
July 2025
Working Paper Number:
CES-25-49
How do advanced technology adoption and venture capital (VC) funding impact employment and growth? An analysis of data from the US Census Bureau suggests that while both advanced technology use and VC funding matter on their own for firm outcomes, their joint presence is most strongly correlated with higher employment levels. VC presence is linked with a high increase in employment, though primarily among a limited subset of firms. In contrast, technology adoption is associated with a smaller rise in employment, yet it influences a considerably larger number of firms. A model of startups is created, focusing on decisions to use advanced technology and seek VC funding. The model is compared with firm-level data on employment, advanced technology use, and VC investment. Several thought experiments are conducted using the model. Some experiments assess the importance of advanced technology and VC in the economy. Others examine the reallocation effects across firms with different technology choices and funding sources in response to shifts in taxes and subsidies.
View Full
Paper PDF
-
Investments under Risk: Evidence from Hurricane Strikes
June 2025
Working Paper Number:
CES-25-43
We demonstrate that firms with plants in areas subject to a significant hurricane strike reduce their capital expenditures at the hurricane-affected plants and shift capital expenditures to plants in non-hurricane-affected areas. This effect is not present prior to 1997 and only appears from 1997 on. Our evidence is consistent with the possibility that a significant climate event such as the signing of the Kyoto Protocol raised the salience of the perceived risk from actual hurricane strikes and shifted firm behavior.
View Full
Paper PDF
-
Dynamics of High-Growth Young Firms and the Role of Venture Capitalists
June 2025
Working Paper Number:
CES-25-38
Motivated by the substantial growth and upfront investments of venture capital (VC) backed firms observed in administrative US Census data, this paper develops a firm dynamics model over the life cycle. In the model, startups choose the source of financing from VC, Angel investors, or banks, depending on their growth potential, and invest in innovation. The calibrated model explains the life-cycle dynamics of firms with different sources of financing and implies that venture capitalists' advice accounts for around 22% of the growth of VC-backed firms. A counterfactual economy without VC financing would lose aggregate consumption by around 0.4%.
View Full
Paper PDF
