In the dissertation, an original methodology was developed for conducting seismic stress tests for precast reinforced concrete buildings on the basis of target seismic risk. The methodology includes both seismic risk analysis and a grading system, which is based on a comparison of the assessed risk with different risk acceptance boundaries. Based on the grades obtained in the grading process, the outcome of the stress test is determined. The possible outcomes are: "pass" (grades AA and A), "partly pass" (grade B), and "fail" (grade C). The stress test can be conducted at two different reliability levels, which differ in the number of seismic response analyses required. The stress test at the higher reliability level was applied to three single-storey precast reinforced concrete buildings with inadequate connections and different types of non-structural elements (vertical panels, horizontal panels, and masonry infills). The outcome of the stress test was negative in all cases, the worst performance being observed in the case of the building with vertical panels. This building was also used in the stress test at the lower reliability level. In this case, less than one percent of all the analyses required at the higher level were conducted; however, the results obtained at the two reliability levels were practically the same. Additionally, seismic stress tests can also be used at the level of systems of buildings, if appropriate tools are available. To this end, fragility functions for twelve building classes, which are one of the inputs for the stress test at the systemic level, were derived. Such fragility functions also make possible comparisons between the seismic responses of different building types. Existing single-storey precast reinforced concrete buildings with inadequate seismic design were examined. It was found that non-structural elements increased seismic risk in the considered buildings by reducing the median collapse intensity and increasing its dispersion.