Seminar of Trade Economists in Paris (STEP)

Supply-chain disruptions are becoming increasingly frequent and severe. To mitigate their vulnerability to these shocks, firms often resort to input stockpiling and supply source diversification. In this paper, I propose a theory where firms choose how many suppliers to source from and how much inventory to hold to maximize expected profits under supply uncertainty. Diversifying supply sources reduces the variance of supply disruptions but involves sunk costs, while inventories reduce the variance of input availability and involve linear costs. I test the key predictions of the model using a comprehensive panel of French manufacturing firms. First, inventories and diversification are substitutes: there is a negative correlation between the inventory ratio — the fraction of inputs held in inventory — and diversification. Second, smaller firms tend to hold more inventory and diversify less, whereas larger firms hold fewer inventories and diversify more. I calibrate the model to examine how inventories help buffer supply shocks and assess the extent to which both strategies enhance firm performance. Absent diversification and inventories, output volatility would be 12.5% higher. Preliminary results suggest that investing in inventories is more cost-effective than increasing supplier diversification.

This paper shows that direct air connectivity promotes and shapes cross-border innovation collaborations. Using unique data on new direct air links connecting Chinese cities to the rest of the world between 1990 and 2014, we find that the introduction of a direct link between a Chinese city and a foreign city increases the number of collaborations between them by 0.5 patents (half a standard deviation of no. of patents) after 10 years on average. The impact does not seem to depend on the number of flights once the air connection is effective, which is consistent with a constraint-unlocking effect rather than variable-cost effect. The increase in the number of patents after direct air connection is only observed when the overlap in working hours between the two connected cities is less than 2 hours. There is no impact as a result of connecting Chinese cities with Asia-Pacific locations with which the overlap in working hours is greater. The impact is stronger for technology classes that are more complex and that involve lab experiments, for which more frequent and in-person interactions may be required, which confirms that cross-border communication costs constitute an important friction to collaboration between inventors.  Overall, results suggest that frictions in synchronous communication are a major barrier to innovation collaboration. Direct flights help only when the constraint is temporal distance and incapacity to communicate and work synchronously.

This paper investigates how firms adapt their sourcing of clean and dirty inputs in response to changes in climate policy. We use information from the European Union’s Emissions Trading System (EU ETS) and the Carbon Border Adjustment Mechanism (CBAM) to create a new classification of clean and dirty products based on whether they are subject to a domestic or a border carbon tax. We then combine this dataset with French firms’ product-level import data over 2000–2019 and estimate that firms’ propensity to import dirty inputs from non-EU countries increased in the 2010s, reflecting carbon leakage. A heterogeneous firm model is then used to quantify the impact of changes in firms’ sourcing of clean and dirty inputs given the implementation of a carbon tax and a carbon tariff. The simulated ETS carbon tax scenario is able to match leakage observed in the data and leads to a higher price level and a modest decline in emissions. The scenario that further includes the CBAM carbon tariff reverses carbon leakage at the cost of an additional rise in prices. Overall, household welfare declines because the higher costs associated with the carbon policies outweigh the benefits of reduced emissions.

This paper studies the export expansion strategies of firms. We classify new seller-buyer- product combinations into two margins: going wide (introducing new products) and going deep (reaching new buyers for existing products). Going wide is the dominant margin for small firms, and going deep for large firms. Further, firms typically target new products to a single, often new, buyer. We rationalize these facts in a model with expansion costs where the returns to scale for going wide and going deep vary. The findings suggest varying scope and size for firms born in different phases of globalisation.