I have thought that there was some funny business going on with BLS data for a long time. Here you can find out why some of the newer models of recession/depression are far more accurate than previous models. What is being done is to add another variable to the model to remove both false positives and false negatives. The results show something much different than is being reported out by BLS.