We compared 5-min standard deviations (SD) and frequency domain measures of beat-to-beat pulse pressure (PP) variability with those of RR-interval, systolic (SBP) and diastolic (DBP) blood pressure variabilities, and with cross-spectral baroreflex sensitivity (BRS) in a population-based sample of 150 healthy individuals, aged 35–64 years. Beat-to-beat variability of PP was composed of similar frequency components as the other spectral variabilities, and was closely related to SBP variability. The proportion of high frequency (HF) component from overall variability was higher in PP variability than in SBP and DBP variabilities. The low frequency (LF) component and the SD of beat-to-beat PP correlated inversely with BRS (−0·48 and −0·32, respectively; P<0·001 for both). To test a hypothesis that arterial stiffening is associated with increased beat-to-beat oscillation in PP, we examined associations of beat-to-beat PP variability with risk factors of atherosclerosis, i.e. with age, gender, smoking, blood pressure, body mass index, serum lipids, glucose, insulin and homeostasis model assessment of insulin resistance. The SD of beat-to-beat PP variability correlated with age (0·21, P = 0·010), PP (0·31, P<0·001) and body mass index (0·22, P = 0·008). The LF component of PP variability correlated not only with age (0·17, P = 0·041), PP (0·27, P = 0·001) and body mass index (0·22, P = 0·007), but also with serum insulin (0·17, P = 0·042), homeostasis model assessment of insulin resistance (0·18, P = 0·031) and serum triglycerides (0·16, P = 0·048). Our findings suggest that increased beat-to-beat oscillation of PP reflects arterial stiffening and impaired baroreflex function.