Stress is a suspected cause of depression. High cortisol concentration, a biomarker of an activated stress response, has been found in depressed patients. The aim of this study was to determine if a high level of salivary cortisol is a risk factor of depression. In 2007, we enrolled 4467 public employees. Morning and evening salivary cortisol concentration were measured for each participant. Participants reporting high levels of depressive, burnout, or stress symptoms, assessed by questionnaires were assigned to a psychiatric interview. In this interview 98 participants were diagnosed with depression and subsequently excluded. Two years later in 2009, 2920 participants who had provided at least one valid saliva cortisol measurement at baseline participated at follow up. The psychiatric interviews were repeated and 62 cases of newly onset depression were diagnosed. Odds ratios of depression were estimated for every 1.0nmol/l increase in morning, evening, and daily mean cortisol concentration, as well as for the difference between morning and evening cortisol concentration. The risk of depression decreased by increasing daily mean cortisol concentration and by increasing difference between morning and evening concentrations, while morning and evening cortisol concentrations were not significantly associated with depression. The adjusted odds ratios for 1.0nmol/l increase in morning, evening, and daily mean cortisol concentration were 0.69 (95% CI: 0.45, 1.05), 0.87 (95% CI: 0.59, 1.28), and 0.53 (95% CI: 0.32, 0.90), respectively. The adjusted odds ratio for 1.0nmol/l increase in difference between morning and evening concentration were 0.64 (95% CI: 0.45, 0.90). This study did not support the hypothesis that high salivary cortisol concentration is a risk factor of depression, but indicate that low mean salivary cortisol concentration and a small difference between morning and evening cortisol concentration may be risk factors of depression.