Some people believe that the impact of population ageing on future health care expenditures will be quite moderate due to the high costs of dying. If not age per se but proximity to death determines the bulk of expenditures, a shift in the mortality risk to higher ages will not affect lifetime health care expenditures as death occurs only once in every life. We attempt to take this effect into account when we calculate the demographic impact on health care expenditures in Germany. From a Swiss data set, we derive age-expenditure profiles for both genders, separately for persons in their last 4 years of life and for survivors, which we apply to the projections of the age structure and mortality rates for the German population between 2002 and 2050 as published by the Statistische Bundesamt. In the extreme case we assume that morbidity is compressed at the end of life in such a way that a 60-year old in 2050 is as healthy as a 56-year old today if his life expectancy is 4 years higher. We calculate that at constant prices, per-capita health expenditures of Social Health Insurance would rise from €2596 in 2002 to between € 2959 and € 3102 in 2050 when only the age structure of the population changes and everything else remains constant at the present level, and to between €5232 and €5485 with a technology-driven exogenous cost increase of 1 % per annum. A "naive" projection based only on the age distribution of health care expenditures, but not distinguishing between survivors and decedence, yield values of €3217 and €5688 for 2050, respectively. Thus, the error of excluding the "costs of dying" effect is small compared with the error of underestimating the financial medical technology.