We are attempting to duplicate the 2007 Mortality tables, which were published in the "Updated Mortality Tables for Determining Current Liability" regulations around 2/2/2007. We are matching qx values for most ages but are slightly off for ages between 41 and 49 and between 71 and 79, which is where the Non-annuitant and Annuitant tables are smoothed. Has anyone been able to match these factors?

An example may help:

Male age 76
qx from published table =.034523
qx per our calculation below =.034489

Non-annuitant qx (using smoothed transition since no actual non-annuitant qx at age 76)
.030710 = .009922 + (.064368 - .009922) x 21/55 (smooth transition discussed in proposed regs on Mortality Tables dated 5/29/2007)
(.064368 is age 80 annuitant qx and .009922 is age 70 non-annuitant qx from published table)

Annuitant qx
.042169 from published table

Non-annuitant qx (projected to 2007)
.022520 = .030710 x (1-.014)^22 (.014 is from projection Scale AA for 76 year old male, 22 years from 2000 to 2007 plus 15)
Annuitant qx (projected to 2007)
.034615 = .042169 x (1-.014)^14 (.014 is from projection Scale AA for 76 year old male, 14 years from 2000 to 2007 plus 7)

Calculated qx formula
.034489 = .022520 x (1-.9896) + .034615 x .9896
(98.96% of those age 76 are annuitants using weighting factors from "Updated Mortality Tables for Determining Current Liability" proposed regulations on 12/2/2005)

Any help would be greatly appreciated.

Sorry, I also meant to upload this file as well.

Doug,
Thanks for the information. We realized that we were smoothing the 2000 rates instead of the projection year rates. With that change we match the published values.
Thanks,
Rolf