Purpose – The purpose of this paper is to investigate whether employers’ attitudes towards older workers, especially regarding promotions, really affect their retirement intentions, distinguishing between men and women. Design/methodology/approach – First, the author uses the 1992 wave of the Health and Retirement Study to estimate, through a Fields decomposition, the relative contribution of the feeling of an older worker to be discriminated against regarding promotions; and to explain the self‐reported probability to work full time after 62, decomposing by gender. Second, using the two first waves of HRS, the author removes any bias due to time‐constant unobserved heterogeneity, to test whether the individual feeling of being passed over for promotion may be misreported, owing to a strong preference for leisure. Finally, the author examines the effect of a change in this variable over time on the intentions to exit early. Findings – The Fields decomposition shows that feeling passed over for promotion plays a non‐negligible role to predict retirement plans but only for women. In addition, using panel data allows a misreporting bias to be exhibited that may lead to underestimating of the negative effect of discriminatory practices towards older workers on their retirement plans. Lastly, an increase between 1992 and 1994 in the age‐discrimination towards older workers encouraged women to leave their job early, while it had no effect on retirement plans of men. Practical implications – Empirical results put forward the idea that retirement intentions may differ across gender, owing to the different nature of the employer‐employee relation. While for men, this relation is characterized by delayed‐payment arrangements signed ex ante with the employer, as already shown by Adams, it is not true for women. Consequently, the age‐based preference of employers for promotion, leading to a lower probability of promotion for older workers, is treated by men as a consequence of ex ante arrangements and does not affect their retirement plans. However, women can attribute such attitudes of their employer to a kind of blatant discrimination, reducing therefore their attachment to their job. Originality/value – The paper presents a longitudinal approach towards the determinants of retirement intentions that allows the unobserved heterogeneity constant over time to be removed and to estimate to what extent the feeling of being passed over for promotion may be attributed, for each gender, to some arrangements signed ex ante with the employer.
