Offshoring is a recurring issue in the public debate on globalization but remains an economic phenomenon that is difficult to quantify. This article looks at the dynamics of offshoring from 1995 to 2017, a long period during which deindustrialization, stabilization and then a revival of industrial employment in France have succeeded one another. It proposes a renewed and unified methodological framework based on the construction of detection model based on machine learning algorithms.

Estimates show a slowdown in the number of relocations (-25% between the annual average for the period 1995-2005 and that for the period 2010-2017) since the 2008 crisis. On average over the period 1995-2017, around 1,000 French companies offshored their activities each year, corresponding to 25,000 jobs per year. The majority of offshoring appear to be mainly industrial, with almost half of them going to European countries. Skilled jobs in industry, including skilled workers, are over-represented among offshored jobs.