Working Papers
AI Unbound: Digital Infrastructure, AI Adoption, and Firm Performance
with Gianmarco Ottaviano
We study how digital infrastructure relaxes constraints on the diffusion and economic impact of artificial intelligence (AI). Using administrative data and a nationally representative enterprise survey from Turkey (2021–2024), we document significant disparities in AI adoption. Adoption is concentrated among large firms and in regions with high-speed broadband and proximity to data centers, particularly for software-intensive and cloud-based applications. To identify causal effects, we exploit the staggered expansion of Turkey’s national natural gas pipeline network, which serves as a conduit for fiber-optic deployment. Because pipeline routing is determined by energy distribution priorities rather than digital demand, it provides plausibly exogenous variation in connectivity. Difference-in-differences estimates show that improved connectivity significantly increases AI adoption, particularly for software-intensive technologies and among small and medium-sized enterprises. Instrumental-variable estimates indicate that infrastructure-driven AI adoption raises labor productivity and export intensity while shifting labor composition toward ICT-related roles. These findings highlight digital infrastructure as a primary determinant of both the pace of AI diffusion and its resulting economic returns.
CEPR Discussion Paper No. 21385 : PDF
Skills, Not Scale GenAI and Technology Adoption
with Gianmarco Ottaviano
Do the determinants of technology adoption depend on technological architecture? Using administrative data on Turkish firms from 2021 to 2024, we compare the adoption of traditional and generative artificial intelligence (GenAI). We show that GenAI adoption is driven by workforce skill intensity and is not positively associated with firm size, whereas traditional AI depends on both scale and skills. Firms that adopt both technologies are distinct and represent the most persistent adoption mode. Conditional on adoption, the skill-to-size ratio governs technology choice, and transition dynamics indicate a sequential process in which firms adopt GenAI before expanding to hybrid use. Exploiting the release of ChatGPT as a quasi-experimental reduction in access costs, we find that high-skill firms differentially increased GenAI adoption, while firm size played a limited role. These results suggest that the canonical size-based diffusion pattern is not universal but depends on the cost structure of technologies, with implications for innovation policy and productivity dispersion.
CEPR Discussion Paper No. 21506 : PDF
Technology Spillovers, Diffusion and Rivalry in Firm Networks
with Ester Faia and Gianmarco Ottaviano
We examine how upstream firms' technology adoption affects the performance and adoption decisions of downstream partners. Using business-to-business data with administrative records on advanced technology adoption, we find gains in productivity, performance, adoption probabilities of firms connected to the adopter, relatively to those that are not. Identification combines staggered event studies, balanced panels of pre-existing relationships, and recentering methods to address expected exposure within the network. Gains vary along firm size, centrality, technology quality, but do not systematically increase with input exposure, suggesting that knowledge spillovers may induce organizational adjustments. Adoption by competitors is associated with short-run negative effects.
CEPR Discussion Paper No. 19804 : PDF | VoxEU Column
Under Review
The Inflationary Cost of Reshoring: Evidence from Firm-to-Firm Networks
How do supply chain disruptions become inflation? Using monthly data on the near-universe of firm-to-firm transactions in Turkey, merged with customs and social security records, I exploit the early COVID-19 lockdown in China as an exogenous disruption to established supply relationships. Firms exposed to the disruption through their pre-lockdown reliance on Chinese suppliers raised output prices by 9.0 log points in the initial months and by 26.9 log points in the long run; differential exposure to the concurrent lira depreciation and global freight surge accounts for well under half of the long-run estimate even under extreme assumptions. The mechanism is forced reshoring: disrupted firms expanded their domestic supplier base by roughly 45.5% while their count of alternative foreign suppliers rose by only 3%, and the substitution toward less efficient domestic suppliers raised marginal costs persistently. A model of endogenous supplier network formation, in which disruptions raise relationship-specific fixed costs and ad valorem tariffs are isomorphic to such shocks on the importing margin, rationalizes the evidence and quantifies the inflationary cost of protectionism. Tariffs operate as self-imposed disruptions, with a 25 percent tariff on the dominant hub priced at roughly 2.2 log points of permanent producer-price inflation. Moreover, domestic substitution saturates once local capacity is exhausted and the aggregate price index rises permanently. Raising domestic supplier productivity delivers resilience at a lower inflationary cost than broad tariffs.
Production Networks and Propagation of Supply and Demand Shocks
with Kamil Yilmaz
PDF
Trademarks and Expansion in Production Networks
with Nevine El-Mallakh
R&R Review of World Economics
Environmental Regulations, Selection and Trade
with Francesco Devicienti, Elena Grinza, Alessandro Manello and Davide Vannoni
SSRN
Circular Economy Investments, MNEs, and Internationalization: Firm-Level Evidence from Italy
with Francesco Devicienti, Elena Grinza, Alessandro Manello and Davide Vannoni
R&R Journal of Cleaner Production
The green transition is a critical challenge for sustainable development, and a better understanding of the determinants of firms’ investments in circular economy (CE) practices is crucial. In this paper, we examine how multinational enterprises (MNEs) and global value chains (GVCs) are associated with sustainabletransformations in production. Using detailed micro-level data on Italian firms combining a representative
firm-level survey with balance-sheet records, ownership data from ORBIS, and the OECD Environmental Policy Stringency (EPS) index. We provide novel evidence on how different forms of internationalisation shape CE and green investment decisions. CE adoption is the primary outcome of interest; green investment is examined as a broader benchmark capturing related environmental upgrading. We show that Italian MNEs (outward FDI) are significant drivers of CE investments. Foreign subsidiaries (inward FDI) invest comparably to domestic firms only when their parent companies are headquartered in countries with stringent environmental regulations, providing evidence consistent with “imported regulation” through multinational networks. Beyond ownership, participation in GVCs further amplifies CE adoption, with effects increasing along a governance gradient: relational and hierarchical GVC modes, which involve closer coordination and knowledge exchange, generate significantly stronger CE premiums than arm-length or quasi-hierarchical ties. These results survive propensity score matching, reducing concerns that observable self-selection drives the findings. The findings highlight the role of institutional context and the quality of international linkages, not merely their existence, as drivers of circular economy adoption.
Publications
Inflation Diffusion through Supply Chains
International Economics, December 2025 : Paper
Policy
Country Economic Memorandum: Leveraging Global Value Chains for Growth in Turkey
World Bank : Report