forcing annual s.a. to equal annual nsa at initial stages

[CHMcLaren]

Not that I'm a fan of this approach, but I couldn't see how to force annual SA to be annual non-seasonally adjusted at the earlier stages of the outputs. I guess it could require an iterative approach. So while I can use benchmarking tab after the SA is finished - this does not take the "forced" SA output all the way through the program. So the calculated trend etc. are not applied on the benchmarked SA, so are then inconsistent (obviously most cases differences are minimal). Is it possible to have an option where the benchmarked SA is consistent with later derived outputs.

[palatej]

Hello Craig,

We didn't invest much in SA benchmarking because it is not recommended in the guidelines of Eurostat.
Most of the benchmarking issues appear in multiplicative decompositions (and model-based methods, which uses rough bias correction). We should investigate some more specialized solutions (see for instance some papers of Tucker McElroy on the topic), not currently implemented. Any technical proposal is welcome.

Jean
rrs2008-01.pdf