Can a one-time, permanent change in the fundamentals behind the sectoral co
mposition of the economy prompt an aggregate downturn? Can this downturn be
nonnegligible, even if one uses US data to determine the relative size of
gross vs. net job flows, and the importance of job creation costs? Can one
consider the military build-down of the 1990s as a plausible cause for the
1990-1991 recession? Do sectoral reallocations generate responses that are
qualitatively similar to 'productivity shocks'? We use a variant of the Mor
tensen-Pissarides (1994, Review of Economic Studies 61, 397-415) job creati
on/destruction model, calibrate it to US labor market data, and run experim
ents that suggest one can answer yes to all these questions. (C) 2000 Elsev
ier Science B.V. All rights reserved.