Scholars generally agree that the number and percentage of bicameral legislatures have been in decline but by the late 1990s the trend has either flattened out or reversed. The existing studies are also predominantly informed by bicameral reforms in mature democracies. Our aim is twofold. First, drawing from new data, we reexamine the fortunes of bicameralism in all democracies and dictatorships from 1945- 2016. Second, we look under the hood of the observed trend and take stock of over a hundred of cases of the introduction, removal, or reintroduction of second chambers. We find that that democracies rarely reform bicameralism and when they do it is typically in context of a transition period. By contrast, the change in dictatorships is much more frequent. Second chambers provide authoritarian leaders with patronage opportunities, which in turn may influence regime stability. Drawing from survival analysis, we find that dictatorships with partly or fully appointed senates are more durable. In general, we argue that it is important to bring the experience of change in non-democratic regimes to discussion about second chamber reform in general