Publications

• Localized effects of confiscated and re-allocated real estate mafia assets, Journal of Economic Geography
  Coauthors: Marco  di Cataldo and Elisabetta Pietrostefani
  In an effort to tackle criminal groups, the Italian State allows the confiscation of properties belonging to individuals convicted for mafia-related crimes and their reallocation to a new use. This policy is considered both as an anti-mafia measure and as a way to partially compensate the society for the harm made by the criminal organizations. Whether and how this measure has had an impact on the local areas where it is implemented, however, has not yet been investigated. We test the hypothesis that the policy contributes to the regeneration of urban spaces by assessing its impact on the value of buildings in the vicinity of confiscated/re-allocated properties. To this aim, we perform difference-in-differences analyses, both at the level of local housing markets and at the level of individual buildings, investigating the externalities of the policy across the whole Italian territory. The results unveil a positive and significant effect of re-allocations of confiscated real estate assets on house prices, declining with distance from the re-allocation site. The impact is larger in cities with stronger mafia presence and in more deprived neighbourhoods. This suggests that the policy contributes to add value to the territory where it is applied and favours processes of urban revitalization. These findings have important implications for the development of deprived urban areas characterized by a strong presence of criminal organizations.
  Final version

Working Papers:

• Capital-Skill Complementarity in Firms and in the Aggregate Economy
  Coauthors:  Giuseppe Berlingieri,  Danial Lashkari, and Jonathan Vogel
  We study capital-skill complementarity in a multi-sector theory of aggregate technology featuring firms with arbitrary and heterogeneous multi-factor production functions. Importantly, the theory allows for distortions to factor choices and heterogeneous skill premia across firms. We characterize the elasticity of the aggregate skill premium to the price of equipment capital in terms of firm-level elasticities of substitution across factors, the elasticities of substitution across firms and sectors, and the distribution of factor intensities across firms. Using micro data from France, we estimate the two sets of elasticities assuming a firm-level production function featuring heterogeneous cross-factor elasticities. Applying our aggregation, we find that a 1% fall in equipment capital prices raises the observed skill premium by 0.06%. This response is mostly driven by the rise in aggregate skill demand due to market reallocations toward equipment- and skill-intensive firms.
  Current version

• High-speed Broadband, School Closures and Educational Achievements (Revise and resubmit)
  In this study, I shed new light on the short-run effects of access to high-speed internet on educational disparities, before and after the pandemic shock. By following 3 million students belonging to 6 different cohorts over the period 2012-2022, I estimate the effect of the broadband infrastructure on student performance. While most previous contributions use discontinuous jumps in the available broadband connection speed across space at a given moment in time, this study exploits the actual roll-out of an infrastructural policy associated with an increase in 30 Mbit/s household broadband coverage from 40% to 80% over a 5-year period. The estimation strategy relies on a unique dataset, combining panel data on student performance with a rich set of school- and student-level information and broadband data measured at a very fine spatial scale. Results show an average null effect of high-speed broadband on 8th grade student performance in both numeracy and maths. However, this results masks substantial heterogeneity: lower performers in grade 5 and students with better backgrounds gain from internet speed, whereas the opposite is true for other students. Interestingly, the stronger effect on low-performers tends to disappear during the lockdown, suggesting a negligible mitigating role for high-speed internet during the period of school closure. On the other hand, the broadband infrastructure might have further amplified the gap between students with different socioeconomic background.
  Geography and Environment Discussion Paper Series (38). London School of Economics

• Offshoring and Local Labour Markets: Evidence from Matched Employer-Employee Data
  Offshoring is often associated with the destruction of low-skilled jobs and the rise of wage disparities. Some firm-level studies have analysed the effect of outward foreign direct investment on wage dispersion. However, little evidence has been provided regarding local spillovers. This article investigates trade shock externalities on local labour markets. Exploiting a unique matched employer-employee dataset, covering the whole population of private sector workers, I link firms’ decision to offshore to worker-level outcomes in the neighbourhood of the establishments directly affected by the shock. Besides the main effects, I explore treatment heterogeneity with regard to worker task content and occupational mobility patterns and I characterize the heterogeneous outcomes in terms of magnitude and spatial decay. At the firm level, results suggest that offshoring raises wages for non-routine and interactive occupations and decreases it for offshorable workforce. Firm level-outcomes are found to impact workers located in the nearby of treated establishments up to 20km distance. For occupation pairs characterized by a high degree of occupational mobility, cross-occupation spillovers play a significant role.
  (available upon request)

Work-in-progress:

• Natural disasters, automation and reshoring dynamics
  Coauthor: Riccardo Crescenzi
  This paper investigates the nexus between reshoring and automation. By mapping the propagation of local shocks across regions and industries trough international ownership networks, we first study how business groups’ (BG) internal organisation reacts to exogenous shocks. International ownership networks are identified through an iterative process, able to detect pyramidal structures and cross-holding exchanges among companies. Applying this methodology on BvD Orbis Historical data, we construct a longitudinal ownership network over the period 2010-2019. This allows us to document to what extent exogenous shocks disrupt firm-level trade with affected regions, foster the divestment of local affiliates and change the overall structure of business groups. Then, we analyse the effect of the idiosyncratic shocks on domestic employment, exploiting a unique matched employer-employee dataset, covering the whole population of French workers. The effect of exogenous divestment decisions is analysed with respect to changes in firm skill composition and investments in domestic automation. As expected, BGs react differently to local shocks depending on the position of the affected affiliate in the organization. Foreign affiliates directly exposed to the shock are more likely to be divested the longer the physical and the ownership distance from the parent firm. Contrary to the previous literature, we find a positive effect of reshoring on both domestic employment and equipment capital. However, this result masks a significant heterogenity across different occupational groups. The effect is stronger for high-skilled occupations and is likely to foster within-firm wage inequalities. Firms hit by network shocks are significantly more likely to invest in industrial robots and in other types of automation capital. These results are consistent with a partial substitution of foreign low-skilled labour with domestic automation.

• Marshall at the times of Marshall
  Coauthors: Olmo Silva
  Was Alfred Marshall a good observer or a visionary? In the late XIX century, the British economist theorised about the existence of three sources of agglomeration economies: labour pooling, input sharing, and knowledge spillovers. Over the last two decades, an extensive literature has tested the existence of the three Marshallian forces in modern economies. However - besides the detailed descriptive case-studies provided by the author himself - there is limited quantitative evidence regarding the existence of such forces at the times of Marshall. To shed light on these issues, we exploit novel geo-localised census-level data on entrepreneurs and business proprietors retrieved from six consecutive UK Censues (1851, 1861, 1881, 1891, 1901, 1911), coupled with census-level workers’ data, information on patents recorded between 1851 and 1911 and historical IO tables. We use these data to estimate co-agglomeration models that estimate the importance of labour pooling, input sharing and knowledge spillovers in explaining clustering of economic activities at the times of British rapid industrialisation. Our preliminary results highlight a significant role for knowledge spillovers and labour pooling, but only a limited effect for the input sharing. We also investigate the location patterns and co-agglomeration drivers that characterised three traditional industries discussed by Marshall, namely: cutlery, textile and machinery. Interestingly, the relative strength of agglomeration is generally consistent with what has been found for modern economies - and to a large extent in line with Marshall’s predictions. Our evidence on the historical importance of the ’Marhsallian trinity’ coupled with the richness of our data lays the foundations for an ambitious research agendas on the spatial evolution of economic activities in late XIX century England, and on the rise and fall of sectors and regions.

• Work-from-home and Firms’ Resilience: Evidence from the COVID-19 pandemic
  Coauthors: Riccardo Crescenzi, Davide Rigo
  By using unique administrative data on firms' adoption of work-from-home (WFH) for Italy, we study the impact of WFH in the response of firms to the COVID-19 shock. Using a difference-in-differences framework and an instrumental variable approach, we examine firms' performance before and after the start of the COVID-19 pandemic by comparing firms' adopting or not WFH  practices during the lockdowns imposed in 2020. We find that the adoption of WFH during the outbreak had a mixed impact on firms, with smaller firms that experienced negative sales growth. Moreover, firms that invested more in information and communication technologies before the start of the pandemic were able to mitigate this negative effect.