Purpose – Applied research shows higher wages are associated with lower mobility at the establishment level. A usual interpretation is that high pay decreases labour turnover. The purpose of this paper is to test if such relationship holds for every type of worker in every type of firm. Design/methodology/approach – The analysis is based on a linked employer‐employee panel dataset covering the French private sector from 2002 to 2005. The authors compute establishment wage effects and use them as explanatory variables in labour mobility equations (for churning rate and quit rate). Using spline regression models enables to investigate for potential non‐linearities. Findings – The authors show that the relationship between churning rate and wage is non‐linear and has the shape of an inverted J: the relation is negative and intense for establishments with low wage effect, weaker for average paying establishments and even becomes positive for very high‐paying ones. This is true whatever the skill group of workers. It is also true for large establishments while the relationship is still negative but linear for small ones. The relationship between wages and quit rates has a strikingly similar pattern. This suggests that the link between churning and establishment wage effect is strongly related to quit decisions. Practical implications – A possible interpretation of our results is that paying higher wages may be an effective stabilizing tool especially for employers in small establishments and when starting wages are relatively low. Originality/value – The paper is the first to decompose the relationship between wage and mobility. It shows the relationship differs across establishment size and is not linear. The paper also shows quits play a role in this relationship.