Risk homeostasis theory is a behavioural theory of risk taking in road traffic. So far, most of the published papers concerning RHT and long time series are based on econometric methods which are not very well suited for this purpose. We propose here to address the issue using the concept of stationarity instead of the econometric methods. We then test the RHT with German traffic accident data and specically analyze compulsory traffic safety measures (the penalty for not using seat belts) which are ineffective according to RHT. Our results, found by using several risk measures, show only weak evidence for RHT. We found that, due to these safety regulations, accident risk was displaced from more to less severe accidents, as the number of severe accidents with fatalities and injuries decreased, as did the average severity of accidents with personal damages. Contrary to RHT, we can show that compulsory safety measures combined with penalties had a strict positive effect on the road traffic accident risk.