Productivity performance, distance to frontier and AI innovation : Firm-level evidence from Europe
(2024) In Journal of Economic Behavior and Organization 228.- Abstract
Using firm-level data from 15 European countries between 2011 and 2019, this paper examines the productivity effect associated with the development of Artificial Intelligence (AI), measured by patenting success in AI fields. By making advances in AI and expanding on their knowledge base, companies can optimise production tasks, and improve resource utilisation, ultimately leading to higher levels of efficiency. To investigate this, we develop a two-fold panel regression analysis estimated within a Difference-in-Differences (DiD) framework. First, we investigate whether firms that engage in AI innovation experience a productivity boost after developing the new technology, compared to similar firms which do not undertake AI innovation. To... (More)
Using firm-level data from 15 European countries between 2011 and 2019, this paper examines the productivity effect associated with the development of Artificial Intelligence (AI), measured by patenting success in AI fields. By making advances in AI and expanding on their knowledge base, companies can optimise production tasks, and improve resource utilisation, ultimately leading to higher levels of efficiency. To investigate this, we develop a two-fold panel regression analysis estimated within a Difference-in-Differences (DiD) framework. First, we investigate whether firms that engage in AI innovation experience a productivity boost after developing the new technology, compared to similar firms which do not undertake AI innovation. To analyse this, we employ a novel event-analysis methodology that quantifies the effect of the treatment (AI innovation) on firm performance (productivity) using a Local Projections approach within the DiD setting. Second, we utilise a Distance-to-Frontier (DTF) regression framework in order to examine whether the productivity premium of AI is associated with a firm's ability to absorb knowledge and learn from the technologies developed by market leaders. Our findings reveal that the productivity gains directly associated with AI are statistically significant and quantitatively important, ranging between 6.2 and 17% in the event analysis, and between 2.1 and 6% in the DTF framework. We also provide some evidence that the productivity benefits of AI might be greater for those firms further away from the frontier (between 0.3 and 0.7%). Our research demonstrates that Artificial Intelligence can play a crucial role in enhancing firm productivity in Europe, a result that is evident even in these early stages of the technology's life cycle.
(Less)
- author
- Marioni, Larissa da Silva ; Rincon-Aznar, Ana and Venturini, Francesco LU
- organization
- publishing date
- 2024-12
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Artificial Intelligence, Distance-to-Frontier, European firms, Local Projections Difference-in-Differences, Multi-Factor Productivity
- in
- Journal of Economic Behavior and Organization
- volume
- 228
- article number
- 106762
- publisher
- Elsevier
- external identifiers
-
- scopus:85206886281
- ISSN
- 0167-2681
- DOI
- 10.1016/j.jebo.2024.106762
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2024 The Authors
- id
- 46c359fb-0a5a-40e5-b4dc-b7d9eaaf365e
- date added to LUP
- 2024-11-26 10:12:38
- date last changed
- 2025-04-04 15:23:16
@article{46c359fb-0a5a-40e5-b4dc-b7d9eaaf365e,
abstract = {{<p>Using firm-level data from 15 European countries between 2011 and 2019, this paper examines the productivity effect associated with the development of Artificial Intelligence (AI), measured by patenting success in AI fields. By making advances in AI and expanding on their knowledge base, companies can optimise production tasks, and improve resource utilisation, ultimately leading to higher levels of efficiency. To investigate this, we develop a two-fold panel regression analysis estimated within a Difference-in-Differences (DiD) framework. First, we investigate whether firms that engage in AI innovation experience a productivity boost after developing the new technology, compared to similar firms which do not undertake AI innovation. To analyse this, we employ a novel event-analysis methodology that quantifies the effect of the treatment (AI innovation) on firm performance (productivity) using a Local Projections approach within the DiD setting. Second, we utilise a Distance-to-Frontier (DTF) regression framework in order to examine whether the productivity premium of AI is associated with a firm's ability to absorb knowledge and learn from the technologies developed by market leaders. Our findings reveal that the productivity gains directly associated with AI are statistically significant and quantitatively important, ranging between 6.2 and 17% in the event analysis, and between 2.1 and 6% in the DTF framework. We also provide some evidence that the productivity benefits of AI might be greater for those firms further away from the frontier (between 0.3 and 0.7%). Our research demonstrates that Artificial Intelligence can play a crucial role in enhancing firm productivity in Europe, a result that is evident even in these early stages of the technology's life cycle.</p>}},
author = {{Marioni, Larissa da Silva and Rincon-Aznar, Ana and Venturini, Francesco}},
issn = {{0167-2681}},
keywords = {{Artificial Intelligence; Distance-to-Frontier; European firms; Local Projections Difference-in-Differences; Multi-Factor Productivity}},
language = {{eng}},
publisher = {{Elsevier}},
series = {{Journal of Economic Behavior and Organization}},
title = {{Productivity performance, distance to frontier and AI innovation : Firm-level evidence from Europe}},
url = {{http://dx.doi.org/10.1016/j.jebo.2024.106762}},
doi = {{10.1016/j.jebo.2024.106762}},
volume = {{228}},
year = {{2024}},
}