Cost-benefit analysis (CBA) is often used when many transport investments need to be ranked against each other, for example in national investment planning. However, results are often questioned on claims that the ranking depends crucially on uncertain assumptions about the future, and on methodologically or ethically contestable trade-offs of different types of benefits relative to each other. This paper explores the robustness of CBA rankings of transport investments with respect to two types of uncertainties: relative benefit valuations and scenario assumptions related to car ownership, characteristics and costs. The study is based on CBAs of 479 suggested road and rail investments in Sweden that have been shortlisted for possible inclusion in the national transport investment plan. The CBA ranking turns out to be robust to variations in the studied scenario assumptions. The CBA ranking also turns out to be robust to changes in the relative valuations of different types of benefits – person travel time savings, traffic safety, emissions and freight benefits. We also compare two sets of travel time valuations against each other, one of which differentiated with respect to mode and travel purpose and one which is not, again concluding that the investment ranking is robust.