Tree retention on clear-cuts is a relatively new measure in forestry aimed at ‘lifeboating’ forest species during young seral periods. However, the effectiveness of tree retention for maintaining biodiversity for more than a few years is still poorly known. We investigated lichen persistence in retained buffer strips along small streams after clear-cutting of the surrounding forest, and compared with clear-cuts and un-cut references. Specifically, we compared richness and frequency of red-listed/signal species, calicioids and pendulous species before clear-cutting with 2.5 years and 16.5 years after clear-cutting, and also analysed their colonization-extinction dynamics over time. The results show that the richness of red-listed/signal species and calicioids in buffer strips had declined significantly after 16.5 years, but not after 2.5 years, while frequency displayed a significant difference already after 2.5 years. The richness of pendulous lichens remained relatively stable over time, but the frequency had declined significantly after 16.5 years. In clear-cuts all groups declined more than in buffer-strips (~2–3.5 times more) and the main decline had occurred already after 2.5 years. References remained stable over time. The colonization-extinctions dynamics reflected the richness declines, with high early extinction in clear-cuts and lower but late extinction in buffer-strips, and low (re)colonization. We conclude that retained buffer strips cannot maintain lichen richness over time due to time-lagged extinction, but they are clearly more effective than clear-cuts. Wider buffer strips could potentially reduce tree mortality and decrease lichen extinction. The large amounts of standing dead wood makes buffer strips potential future colonization targets.