Communication dans un congrès: This paper uses a unique data set of linked employer-employee data in which asset transfers between firms can be identified to study the manner in which employment policy changes in the aftermath of a merger or acquisition (M&A). Using parametric duration models with unobserved heterogeneity and modeling the status of the employer with respect to M&A activity, it appear that the employment policy of the firm changes radically after an M&A with respect to the “steady state”. Not only do various observed characteristics impact the probability of continued employment in a different manner after an M&A, but the distribution of unobserved characteristics that affect employment changes – reflecting differences in the stock of previous employees and the flow of new hires – as does the impact of this heterogeneity on employment durations.