An analysis of panel data on individuals in a random selection of urban households in Ethiopia reveals large, sustained, and unexplained earnings gaps between public and private, and formal and informal sectors over the period 1994-2004. The authors have no formal evidence whether these gaps reflect segmentation of the labor market along either of these divides. In other words, they cannot show whether they are at least in part due to impediments to entry in the higher wage sector. But they do have evidence that, if segmentation explains any part of the observed earnings gaps, then it could only have weakened over the survey decade. The authors find, first, that the rate of mobility increased between the two pairs of sectors. Sample transition rates grew across survey waves, while state dependence in sector choice decreased. Second, the sensitivity of sector choice to earnings gaps increased over the same period. In particular, the role of comparative earnings in selection into the informal sector was evident throughout the survey decade and increased in magnitude over the second half of the period.