The construction bust which accompanied the Great Recession, and the accompanying need to shift workers across sectors, have provoked a discussion about mismatch and the Beveridge Curve, alongside a discussion about firm-level dispersion. These discussions echo an ongoing discussion about the effects of long run sectoral reallocation. Based on estimates from a large state space model over a long sample for the United States, long run sectoral reallocation does not appear to be systematically related to movements in the Beveridge Curve, although reallocation does appear to be countercyclical and related to falls in the trend employment-population ratio. The recent shift in the Beveridge Curve during the Great Recession is unusual in this respect. An analysis of historical patterns reveals a handful of additional reallocative episodes, with large episodes occurring during the mid-1970s and early 2000s recessions, in addition to during the Great Recession. In addition, these episodes appear to be related to other dispersion shocks which have been increasingly discussed in the literature.