This paper investigates the statistical properties of the U.S. sacrifice ratio -- the cumulative output loss arising from a permanent reduction in inflation. We derive estimates of the sacrifice ratio from three structural VAR models and then conduct Monte Carlo simulations to analyze their sampling distribution. While the point estimates of the sacrifice ratio confirm the results reported in earlier studies, we find that the estimates are very imprecise and that the degree of imprecision increases with the complexity of the model used. That is, increases in the number of structural shocks widen our confidence intervals. We conclude that the estimates provide a very unreliable guide for assessing the output cost of a disinflation policy.