FIRMNET – Firms and their Networks

There is a mounting evidence that firms are becoming more fragmented; production is less often made “in-house”. Firms buy inputs from abroad. Tasks are often split in parts. Some are off-shored, others are subcontracted. Hence, firms also buy services from other, local or international firms. They also supply inputs to other firms. Technical change, the internet, and globalization, all facilitate the division of labor between firms in the global economy. In order to study firms’ structure and behavior in this changing world, one must examine the relations these firms have built with their direct environment and how such relations shape firms’ answers to shocks. The project aims to construct a networks view of the firm, addressing empirical, theoretical, and econometric questions.


The research agenda consists of four closely related Work Packages (WP). the first two are empirical and focus on how social networks (WP1, on Sweden) and business networks (WP2, on French exporters and their international clients) help firms to be resilient when facing shocks. Fragmented and specialized firms create a business network of customers and suppliers to secure provision of different tasks. Moreover, firms may rely on their employees’ social networks (family ties, boardroom relations, …) to find new suppliers of good and servives, hire workers or look for administrative or financial support during downturns. Firms’ resilience is essentially mesured using administrative ata on its economic outcomes such as production or value-added and its employees’ labor market outcomes such as employment, wages, or the structure of their occupations.The third, WP3, adds an optimizing dimension together with a system perspective to the questions raised in WP2. It examines a network theory of production and firm-to-firm trade within a General Equilibrium framework, where buyers meet sellers to fulfill production tasks. The model generates both micro- and macro-economic predictions. Finally, inspired by problems faced in WP1 and WP2, WP4 provides econometric tools to understand many-to-one matches (typically n workers and their employing firm or exporter and its n international buyers) in the presence of networks that connect workers or firms. These tools will be applied to questions raised in WP1 and WP2, using the same data sources.

The complete research proposal can be found here.

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant agreement No. 741467)


Francis Kramarz – Principal investigator

Francis Kramarz is a Professor at ENSAE, Research Director there, and visiting Professor at Uppsala University, Sweden. He was the director of CREST (Center for Research in Economics and Statistics), jointly created by École polytechnique and ENSAE (in association with the CNRS) until the end of 2019.

He was an Associate Professor at École polytechnique, invited Professor at Yale, but also a member of the French low-pay commission (groupe d’Experts sur le SMIC), of the French Council of Economic Advisors (Conseil d’Analyse Économique), of the Conseil d’Orientation pour l’Emploi, as well as the Attali commission.

Francis Kramarz now focuses on his research, trying to understand how social networks (family, former co-workers, …), firms’ networks (groups, clients or suppliers) affect firms’ outcomes, in particular their resilience in the face of shocks. For this line of investigation, he received and ERC Advanced Grant.

Beyond his research, Francis Kramarz has written, with Pierre Cahuc “De la précarité à la mobilité : vers une sécurité sociale professionnelle” (2005) and more recently, with Philippe Tibi, in 2016 “Plus de Marché Pour Plus d’Etat” with a foreword by Emmanuel Macron, a book which has receibed the Prix Turgot in 2017.

Francis Kramarz graduated from École polytechnique in 1976, ENSAE in 1981, and holds a PhD (and HDR). He was elected a Fellow of the Econometric Society in 2013 and Fellow of the Society of Labor Economists in 2019.

Personal website

Partnered institutions:

Research Associates:

Research Assistants:

  • Gautier Lenfant
  • Lorenzo Kaaks
  • Sophie Nottmeyer
  • Etienne Guigue
  • Maxime Cugnon
  • Martin Gojani

Work in progress

WP1: Social Networks (Swedish data)

Social Connections and the Sorting of Workers to Firms” (2023), with O. Skans, L. Hensvik and M. Eliason. Journal of Econometrics, 233, 468-506.

The paper was presented on the following occasions:

  • INRA Conference, Rennes, France – May 16, 2019
  • Bank of Italy, Rome, Italy – June 14, 2019
  • AKM Conference, New-York, United States of America – October 12, 2019

Connecting the Young: High School Graduates’s Matching to First Jobs in Booms and Great Recessions” (May 2020), by L. Hensvik, D. Müller and O. Skans. Revised and Resubmitted, 3rd round, Economic Journal.

WP2: French Exporters and their Business Networks

Given the granularity of international trade networks, the present analysis aims to provide empirical evidence on how idiosyncratic, firm-level shocks that affect one side of the trading relation are transmitted to sellers in France, for whom we can measue outcomes using administrative, linked employer-employee data. In particular, we hope to answer the following questions:

  • What happens to French exporters when a (large) buyer disappears?
  • How does the firm redeploy its sales in the face of sudden micro-economic demand shocks?
  • How are workers affected in their jobs or wages? How does the structure of occupations and tasks across and within firms change as a result of global trade?

The yet-to-come results will constitute the first direct measures of the international transmission of shocks on labor market outcomes.

Advancement of the project:

  • We have built an enriched data base on suppliers (French firms) and buyer (European firms), including data on suppliers’ balance sheets, employment and firm-to-firm exports. We also developed a procedure to identify likely buyer deaths from French Customs data.
  • We estimate the effects of “buyer death shocks” on French exporters in an event study design. Our fist estimates show a significant and permanent decrease in export sales of about 20 percent within 5 years following such a death. However, this result also masks a large heterogeneity across firms, which we are currently analyzing. We will then move towards how the shocks translate into job losses or wage movements.
  • We will also try alternative identification strategies, in particular by estimating survival probabilities of foreign buyers in the sample. These probabilities can be used as propensity scores to implement a matching estimator.

WP3: Theory: A General Equilibrium Model of Firm-to-Firm Trade

Pure competition setup:

Firm to Firm Trade: Imports, Exports, and the Labor Market“, with J. Eaton and S. Kortum.

R&R at Econometrica.

The paper was presented on the following occasions:

  • Conference (F. Kramarz) – University of Kent, Canterbury, United Kingdom – November 13, 2018
  • Seminar (S. Kortum) on “Firm-to-Firm Trade” – European University Institute, Florence, Italy – December 13, 2018
  • Graham Lecture (S. Kortum) on “In Search of Trade Frictions” – Princeton, United States of America – April, 18, 2019
  • Keynote (F. Kramarz) – Conference of Baltic economists, Riga, Latvia – June 8, 2019
  • Keynote (S. Kortum)  on “In Search of Trade Frictions” – China Meetings of the Econometric Society (CMES2019), Guanzhou, China – June 19, 2019
  • Seminar (S. Kortum) on “Firm-to-Firm Trade” – University of Tokyo, Tokyo, Japan – June 26, 2019
  • Keynote (S. Kortum) on “In Search of Trade Frictions” – Asia Pacific Trade Seminars (APTS2019), Tokyo, Japan – June 29, 2019

Bertrand competition (work in progress):

We extend the purely competitive framework (see above) to a setting with Bertrand pricing and rent-sharing between workers and firms. This model is able to explain wage differentials between similar workers employed in different firms, while also matching the same aggregate trade features as the pure competition version.

WP4: Economics and Econometrics of Many-to-One Matches

Many-to-one matching between workers and firms:

The goal is to build a game-theoretical model of many-to-one matching and to develop a corresponding estimation strategy.

Skills “bunding” and the uberization of our economies (joint with Philippe Choné): the first working paper is “Matching Workers’ Skills and Firms’ Technologies: From Bundling to Undundling

We try to understand many-one-matching through the prism of the supply of and demand for workers’ skills. Skills are multidimensional and can only be sold as a “bundle”. In the corresponding labor market equilibrium, workers with similar comparative advantage in the k potential skills match to firms with compatible technological preferences in these skills. In today’s world, some og the previously bundled skills are faced with the opening of a market for that exact skill, something we label as “unbundling”. This can be due to technological shocks transforming a market (such as the advent of Uber), legislative changes or the opening to international trade. We characterize how the equilibrium changes, which firms and which types of workers benefit from unbundling and, by contrast, which firms and workers are negatively affected. Ultimately, we want to confront this theory with Swedish data on cognitive and non-cognitive skills.

See also Choné, Gozlan, Kramarz forthcoming SIAM J. on Mathematical Analysis: “Weak Optimal Transport with Unnormalized Kernels“.