(with Andrea Fronzetti Colladon, Barbara Guardabascio, Ludovica Segneri,  Alessandro Sterlacchini)

We investigate the dynamic effects of radical innovation on firm productivity growth and its main transmission channels. Using panel data on over 30,000 Italian firms from 2012 to 2020, and analyzing the text of more than one million patent abstracts, we develop a measure of radical innovation  (novelty) based on the content distance between each patent and the state-of-the-art technological knowledge.  We examine how firm productivity growth responds to the launch of a novelty, finding output per worker to decline after the technological break, but then to rise  in the following years. This finding aligns with the J-curve trajectory, where the early phases of novel technology adoption are marked by large adjustment costs and unmeasured output. The cumulative net effect of radical innovation on productivity growth is positive and economically significant: firms that introduce novel technologies experience, on average, 10\% higher productivity growth than those without such innovations. We find that this effect is primarily driven by employment adjustments. Our results are robust to controlling  for all the most important sources of  productivity growth, for causality issues and other econometric concerns. Notably, the observed pattern is not confined to specific sectors or types of firms, but it consistently holds also among SMEs.


(with Larissa Marioni, Ana Rincon-Aznar)

Using novel data on UK businesses, we identify the key factors that influence firms' propensity to invest in data assets and assess how these investments relate to productivity outcomes. We find that businesses investing in digitised information and in targeted training are 20–40\% more likely to use data for external and scientific purposes, as well as to perform skill-intensive data tasks. Labour productivity levels in data-investing firms are 1.4 to 1.5 times higher compared to firms not engaged in such activities. 


(with Andrea Fronzetti Colladon, Antonio Minnti, Carmelo Parello)

This paper examines the impact of technological interdependence on economic growth in an increasingly integrated world. We construct a Schumpeterian growth model that highlights how innovation not only propels the expansion of leading economies but also facilitates the transfer of knowledge to developing nations. The ability of follower countries to effectively utilize technology transfers depends on their proximity to the technological leader. By catalyzing the technology catch-up of the followers, technological interdependence diminishes the innovation growth potential for the frontier economy. Subsequently, using half-century data from a global sample of countries, we evaluate the predictions of the model by measuring international technology interdependence based on the textual similarity of over 7 million patent applications. Our empirical findings underscore the significance of technological interdependence as a growth driver for technology leaders and, notably, for countries below the frontier. Furthermore, our data reveals a shift in technological interdependence over time, with China emerging as a dominant player in the world's technology market, replacing the United States


(with Carmelo Parello)

This paper develops a growth framework with international knowledge spillovers driven by learning-by-investing externalities that is able to replicate most stylized facts on income convergence and economic growth. The  model predicts that knowledge spillovers from the frontier  enhance the relative levels of productivity and income of the laggards, but  materialise only if the gap in capital intensity   between the frontier and lagging economies  is not too wide. We bring the model to the data and, for a global sample of countries, observe that relative capital does mostly shape their growth pattern, in line with the predictions of our theory. We also document that differentials in income growth are driven by the gaps in capital intensity with respect to the frontier economy (the United States) only below an identified threshold. This effect is independent of knowledge inputs such as human capital and innovation. Our findings suggest that the absorption of knowledge through learning processes and embodied technological change remains a crucial driver of income growth for a significant number of lagging economies worldwide. 



    (with Andrea Bacchiocchi, Germana Giombini, Ludovica Segneri)

Using a panel sample of innovating firms in Italy during the early (pre-LLM) phase of AI development (2012–2021), we examine how financial conditions affect firms’ patenting in new digital fields and, through an event analysis, we assess how these factors influence firm economic performance by shaping AI innovation. Our findings show that higher leverage and larger cash holdings significantly constrain firms’ ability to succeed in new digital fields. Through this channel, productivity is between 1% lower for each additional percentage point in these financial conditions.